perspectives

Edited ty
 A
The nature of creativity

Sitii" AS 133?
H'rfsnai! te*™*#* at HeaBfc
r» . ,
 The nature of creativity
•o cy_- %7

Contemporary psychological
perspectives

Edited by
ROBERT J. STERNBERG
Department of Psychology
Yale University

CAMBRIDGE UNIVERSITY PRESS

Cambridge
New York New Rochelle Melbourne Sydney
Published by the Press Syndicate of the University of Cambridge
The Pitt Building, Trumpington Street, Cambridge CB2 1RP
32 East 57th Street, New York, NY 10022, USA
10 Stamford Road, Oakleigh, Melbourne 3166, Australia

© Cambridge University Press 1988

First published 1988

Printed in the United States of America

Library of Congress Cataloging-in-Publication Data

The nature of creativity.
Includes indexes.
1. Creative ability. I. Sternberg, Robert J.
[DNLM: 1. Creativeness. 2. Psychology. BF 408 N285]
BF408.N354 1988 153.3'5 87-20928

British Library Cataloguing in Publication Data

The Nature of creativity: contemporary
psychological perspectives.
1. Creative ability - Psychological aspects
I. Sternberg, Robert J.
153.3'5 BF410

ISBN 0 521 33036 X hard covers

ISBN 0 521 33892 1 paperback
 Contents

Preface page vii

List of contributors ix

Introduction 1

Part I The role of the environment in creativity

1 The conditions of creativity 11

Beth A. Hennessey and Teresa M. Amabile

Part II The role of the individual in creativity

Psychometric approaches
2 The nature of creativity as manifest in its testing 43
E. Paul Torrance
3 Putting creativity to work 76
Frank Barron
4 Various approaches to and definitions of creativity 99
Calvin W. Taylor

Cognitive approaches
5 A three-facet model of creativity 125
Robert J. Sternberg
6 Problem solving and creativity 148
Robert W. Weisberg
7 A computational model of scientific insight 177
Pat Langley and Randolph Jones
8 Freedom and constraint in creativity 202
Philip N. Johnson-Laird
9 Creativity as a mechanical process 220
Roger C. Schank

v
VI Contents

Part III The role of the individual-environment interaction in

creativity

The study of creative lives

10 Inching our way up Mount Olympus: the evolving-systems

approach to creative thinking 243
Howard E. Gruber and Sara N. Davis
11 Creativity: dreams, insights, and transformations 271
David Henry Feldman
12 Creative lives and creative works: a synthetic scientific approach 298
Howard Gardner

The study of creative systems

13 Society, culture, and person: a systems view of creativity 325

Mihaly Csikszentmihalyi
14 Creativity and talent as learning 340
Herbert J. Walberg
15 The possibility of invention 362
D. N. Perkins
16 Creativity, leadership, and chance 386
Dean Keith Simonton

Part IV Integration and conclusions

17 What do we know about creativity? 429

Twila Z. Tardif and Robert J. Sternberg

Author index 441

Subject index 447
 Preface

Research in creativity has taken on the role of a prodigal stepbrother to research on

intelligence. The work in creativity is a “stepbrother” because although it is in the
same family, it never quite seemed, to many of those who follow it, to measure up.
And indeed, it is the work in creativity that is continually compared to work in
intelligence, rather than the other way around. The stepbrother never seems to be
quite as good or quite as well accepted, regardless of what he does. The stepbrother
is prodigal because of his tendency to go beyond accepted bounds, and even at times
to be grandiose. Whereas intelligence has seemed like a problem that, with suffi¬
cient effort, could be resolved, creativity has sometimes seemed irresoluble, and
some have even wondered if it exists as a single entity or class of entities, aside
from in name. Various approaches have been taken to dealing with this prodigal
stepbrother. Some, like J. P. Guilford, have tried to subsume creativity under the
rubric of intelligence. Others, like Jack Getzels, have argued that creativity is an
entity that is psychologically distinct from intelligence. Still others have simply
ignored the connection between creativity and intelligence. The contributors to this
book represent a diversity of perspectives, but all have in common that they have
studied creativity for its own sake.
Creativity research, which had been stagnant for some time, received a new lease
on life from J. P. Guilford’s 1950 APA presidential address on the nature of
creativity. Guilford brought back into the eye of the psychological public an entity
that had long been languishing. During the 1950s there was a burst of research on
creativity, and it seemed that at least some of the problems of creativity could be
licked. Some of the major contributors during that era of research are represented in
this book. Much of the research of that period was summarized in Philip Vernon’s
1960 book entitled Creativity.
Some of the research on creativity continued into the 1960s and even 1970s, but
other efforts faltered. It became clear by the end of the 1970s that the initial thrust
would slow down and that only a few diehards, such as Howard Gruber and others,
were continuing to study the field in a way that could make a difference. At least
some researchers probably shared my attitude in the late 1970s, which was that few
psychological issues are more interesting, but more intractable, than the issue of
creativity. My remark to my colleagues was that there was no phenomenon that I

vii
viii Preface

would rather study, but that I just did not have any creative ideas regarding how to
study it, and I was not sure how many other people were having creative ideas about
it either.
By the early to middle 1980s, I became convinced that important work on
creativity was once again being done from a number of different perspectives. As I
surveyed the field, I became convinced that a not insubstantial number of the early
investigators were continuing to study creativity in fruitful ways and that some new
investigators were starting to have some genuinely penetrating insights into its
nature. I then started to do a bit of research on creativity, although it did not become
a major focus of my research. I became convinced that a lot of new contributions
had been and were being made, but that they were not all as well known or as well
integrated as they should be because these contributions were scattered into so many
different outlets of publication. My conviction that the field was genuinely re¬
awakening and that there was a need to bring it together motivated this book. I
decided to ask those individuals whom I considered to be the brightest lights in the
field of creativity to contribute chapters to an edited book on the topic, and to my
delight, almost all of them agreed to do so. Their efforts are represented in the
chapters in this book.
I am grateful to the members of my research group who have influenced my
thinking about creativity, particularly Janet Davidson, whose collaboration with me
in studying insight has been particularly influential in my thinking. I am also
grateful to Susan Milmoe, my editor at Cambridge University Press, who has seen
me through yet another book. Most of all, I am grateful to the contributors to this
book for taking time out from their busy schedules to write these chapters. This
book is dedicated to J. P. Guilford, who showed how creativity could be understood
as a part of intelligence, and to Jack Getzels, who showed how it could not be.

Robert J. Sternberg
 Contributors
Teresa M. Amabile
Department of Psychology
Brandeis University
Waltham, MA 02254
Frank Barron
Department of Psychology
University of California
Santa Cruz, CA 95060
Mihaly Csikszentmihalyi
Department of Psychology
University of Chicago
Chicago, IL 60637
Sara N. Davis
Department of Psychology
City University of New York
New York, NY 10036
David Henry Feldman
Department of Child Study
Tufts University
Medford, MA 02155
Howard Gardner
Boston VA Hospital
Boston, MA 02130
Howard E. Gruber
Department of Psychology
University of Geneva
1121 Geneva 4 Switzerland
IX
Beth A. Hennessey
Department of Psychology
Wellesley College
Wellesley, MA 02181
Philip N. Johnson-Laird
MRC Applied Psychology Unit
Cambridge, CB2 2EF England
Randolph Jones
Department of Computer Science
University of California
Irvine, CA 92664
Pat Langley
Department of Computer Science
University of California
Irvine, CA 92664
D. N. Perkins
School of Education
Harvard University
Cambridge, MA 02138
Roger C. Schank
Department of Computer Science
Yale University
New Haven, CT 06520
Dean Keith Simonton
Department of Psychology
University of California
Davis, CA 95616
X Contributors

Robert J. Sternberg E. Paul Torrance

Department of Psychology 185 Riverhill Drive
Yale University Athens, GA 30606
New Haven, CT 06520
Herbert J. Walberg
Twila Z. Tardif College of Education
Department of Psychology University of Illinois
Yale University Chicago, IL 60680
New Haven, CT 06520
Robert W. Weisberg
Calvin W. Taylor Department of Psychology
Department of Psychology Temple University
University of Utah Philadelphia, PA 19122
Salt Lake City, UT 84112
 Introduction

Robert J. Sternberg

The goal of the chapters in this book is to explore the nature of creativity. The
chapter authors take a wide variety of approaches in pursuit of this goal. The book is
divided into four basic parts.
The first part, containing a single chapter, deals with the relationship between
creativity and the external environment of the individual. In their chapter “The
Conditions of Creativity,” Hennessey and Amabile describe what might be viewed
as the necessary prerequisite conditions for creative performance. These conditions
include things such as absence of external rewards and emphasis on intrinsic rather
than extrinsic motivation and performance. Indeed, the authors formulate an intrin¬
sic motivation principle of creativity: People will be most creative when they feel
motivated primarily by interest, enjoyment, satisfaction, and challenge of the work
itself - not by external pressures. This chapter is useful in setting the stage for the
rest of the book because it specifies the environmental constraints that either facili¬
tate or impede manifestation of the creative potential that lies within the individual.
The second part of the book deals with the relationship between creativity and the
internal environment of the individual. This part of the book is divided into two
sections, one focusing on the psychometric approach to creativity and the other
focusing on cognitive approaches.
The psychometric approach is represented in the chapters by Torrance, Barron,
and Taylor. Torrance describes creative thinking as a process of recognizing gaps
and missing elements, formulating hypotheses about what is missing that creates
these gaps, testing these hypotheses, revising and retesting the hypotheses, and,
finally, communicating the results. He considers a number of formal definitions of
creativity and also criteria for recognizing creative performance and then spends the
bulk of his chapter reviewing the literature on the Torrance Tests of Creative
Thinking. These tests measure skills such as fluency, flexibility, originality, elab¬
oration, and so forth. The correlational data regarding the tests are impressive. They
predict quite well, for a number of different groups, criteria such as quality and
quantity of creative performance.
Barron, too, focuses his chapter on a test, but in his case the test is the Symbolic
Equivalence Test. In this test, subjects are asked to think of metaphors, or sym¬
bolically equivalent images, for certain suggested stimulus images. Consider, for

1
2 Introduction

example, “leaves being blown in the wind.” Some possible symbolic equtivalences
would be a civilian population fleeing chaotically in the face of armed aggression,
or handkerchiefs being tossed about in an electric dryer. Different responses are
given differing scores, depending on their quality. The test is scored for number of
acceptable responses and for number of original responses. Barron has found that
his test of creativity yields substantial positive correlations for creative performance
in a variety of domains, including creative writing, architecture, art, and en¬
trepreneurial pursuits.
Taylor focuses his chapter around psychometric conceptions of creativity that
have evolved over a period of years. He describes his own and others’ conceptions,
focusing in particular on his “circular model.” In contrast to Torrance and Barron,
who focus on issues of assessment, Taylor brings the focus toward training, describ¬
ing his “totem method” for developing creativity.
Whereas the psychometric approach concentrates on individual differences in
creativity and their correlates, the cognitive approach concentrates on the mental
processes and structures underlying creativity. The second section of the second
part of the book opens with a chapter by Sternberg, who presents a “three-facet
model of creativity.” According to this model, there are three basic aspects that
interact to generate creative performance. The first aspect is strictly cognitive,
involving those aspects of intelligence having interface with creativity, for example,
processes of insight. Sternberg draws this cognitive facet from his triarchic theory
of human intelligence. The second aspect involves matters of intellectual style, in
other words, how one uses one’s intelligence. Drawing from his theory of intellec¬
tual styles as mental self-government, Sternberg describes, for example, how three
different functions in the use of intelligence are the legislative, executive, and
judicial. The legislative function concerns creation, the executive function involves
acting on creation, and the judicial function has to do with evaluating creation.
Although all three styles can play some role in creativity, Sternberg argues that the
legislative function is the most important and that individuals with a primarily
legislative style are more likely to express themselves creatively than are individuals
with a primarily executive or judicial style. The last aspect of the model pertains to
personality variables. Sternberg argues that certain personality attributes are more
likely than others to lead to creative performance. In particular, individuals with
high tolerance of ambiguity, willingness to surmount obstacles, willingness to
grow, and so on, are more likely to be creative than are others. In sum, creativity is
at the interface among the three aspects of the model.
In his chapter “Problem Solving and Creativity,” Weisberg concentrates on
what could be viewed as a central type of creativity, namely, insight. Perhaps the
most widely held view of insight is what might be called the “special-process”
view - that insight differs in kind from other problem-solving processes. Weisberg
argues that it does not, that, in fact, insight processes are an extension of ordinary
perceptual, memory, and problem-solving processes. He considers some of the
great insights of history and how they can be accounted for in terms of ordinary
Introduction 3

rather than extraordinary information processing. He also recites the results of

several experiments done in his laboratory that argue for his “nothing-special”
view of insight. Weisberg’s research shows, at minimum, thatinsights are not likely
to happen in the absence of a great deal of task-relevant prior knowledge. Whatever
they are, they do not come from nowhere, but rather from the presence and integra¬
tion of the knowledge representation that is usable and useful in a given task.
Whereas the first two chapters in this section derive their theory and base their
research on human subjects, the last three chapters base their theory on research more
in the tradition of artificial intelligence. The first of these chapters, that by Langley
and Jones, presents “A Computational Model of Scientific Insight.” Lang¬
ley and Jones, like Weisberg, focus primarily on insight-related phenomena. Langley
and Jones review three alternative computational models of insight and end up
rejecting all three of them. Whereas the past theories relied primarily on the notion of
a problem space, Langley and Jones, like Weisberg, emphasize memory phenomena.
Also like Weisberg, they especially emphasize the role of analogy in memory.
Insights are especially likely to occur when insightful problem solvers recognize
analogies between new problems they are currently facing and problems they have
solved before. The authors argue that insight is primarily dependent on analogical
retrieval and that this retrieval occurs during the illumination stage of insightful
problem solving. Moreover, the retrieval usually is cued by some external event.
They suggest that a spreading activation model would account for how the external
event touches off the analogy and hence the insight, and they build their theory of
insight around the concept of spreading activation.
A related but somewhat different approach to the problem of creativity is taken by
Johnson-Laird in his chapter “Freedom and Constraint in Creativity.” Johnson-
Laird contends that creativity depends on arbitrary choices and on a mental device for
producing such choices. He believes that a key to the processes of creativity is free
will. Moreover, he likens creativity to a murder: Both depend on a motive, means,
and opportunity! Johnson-Laird has devised a program that simulates creative perfor¬
mance in the domain of music. The program takes a chord sequence as input and
delivers a bass line as output. To produce notes, it will generate a small step, a large
step, or the same note again, according to the rules of the grammar for contours that
Johnson-Laird has derived from the corpus of improvisation. Johnson-Laird notes
that the program is quite competent, but that it lacks two abilities of jazz players: It
makes no specific use of chromatic runs of several passing notes, and it makes no use
of motifs. Johnson-Laird’s chapter is a good example of how ideas about creative
performance can be operationalized through computational mechanisms.
In his chapter “Creativity as a Mechanical Process,” Schank also bases his
approach on artificial-intelligence notions, but applies these notions in a way quite
different from that of Johnson-Laird. Schank, like Weisberg and like Langley and
Jones, relies heavily on notions from human memory in his conceptualizing. In¬
deed, he draws heavily on his dynamic memory model of processing in order to
understand creative performance. He believes that there are basically two sub-
4 Introduction

processes that characterize creativity. The first is a search process, which involves
finding explanations. The second is an alteration process, which modifies and
adapts the explanation in such a way as to allow an explanation originally derived
from one situation to be relevant to another. Notice how this model, like that of
Weisberg, draws heavily on analogy. Indeed, Schank notes that creativity results
from applying old explanations in places where those explanations were never
originally intended to apply. Thus, he believes that creativity depends on two
primary factors: a set of methods for getting reminded, and a set of methods for
adapting remindings in such a way as to fit a new situation. Schank shows how his
notions can be used to generate creative behavior in a fairly wide variety of
situations.
Part three of the book deals with the interaction between the internal and external
environments of the individual. This part of the book is divided into two sections.
The first section contains chapters that use the study of creative lives as a primary
data base; the second section comprises chapters that are more generally historical
in their orientation, using larger data bases rather than studies of small numbers of
lives.
The first chapter in the first section of the third part of the book is by Gruber and
Davis- Their chapter, which describes Gruber’s “Evolving-Systems Approach to
Creative Thinking,” reviews Gruber’s lifetime study of the creative enterprise,.as
well as Davis’s dissertation work. According to the evolving-systems view,
creativity bears virtually no resemblance to the flashes of creative insight that
provide the basis for analysis in some of the earlier chapters. Rather, it is something
that evolves over the course of a lifetime, combining manifold minor insights with
some major ones and directed by a large-scale evolving enterprise. Gruber has
intensively studied Darwin and the development of his theory of evolution and has
shown how the theory of evolution was not a sudden major insight, but rather the
culmination of scores of minor and major insights. Creativity is seen as a slowly
evolving process of reflection and discovery rather than a sudden burst of
enlightenment.
The next chapter is by Feldman, who has been heavily influenced by Gruber.
Like Gruber, he describes the evolution of a set of creative ideas, but in Feldman’s
case, the creative ideas are his own. Whereas Gruber has focused in his research on
adult creativity, Feldman has focused on child prodigies, and he reviews some of
his work on prodigies in this chapter. He shows that some of our common concep¬
tions about prodigious performance are incorrect. For example, many of us believe
that prodigies are so outstanding that their prodigiousness would show under any
environmental circumstances. Feldman shows, to the contrary, that prodigious per¬
formance in children results only because of a rare and complex coincidence of
individual, family, societal, and cultural variables. Without just the right mix, the
prodigious performance will never be achieved.
This aspect of prodigiousness distinguishes it from adult creativity, wherein some
“bucking” of the environment is almost always necessary for creative performance
Introduction 5

to take place. One interesting aspect of prodigious performance is that it is by no

means randomly distributed across fields. Whereas adult creativity can occur in
almost any field, prodigious performance tends to be limited to a fairly small set of
fields, particularly a field in which there is a constrained and clearly defined set of
symbols to manipulate. Thus, chess, based on the manipulation of a small number
of chessmen, is a more likely arena for prodigious performance than writing, which
is based on the manipulation of tens of thousands of words and syntactic forms.
Feldman’s chapter provides a nice complement to Gruber’s chapter in showing
similarities and especially the differences between high levels of creative perfor¬
mance in adults and prodigious performance in children.
Like Gruber and Feldman, Gardner has chosen to study a small number of
creative lives in his attempt to understand creativity through a “Synthetic Scientific
Approach. ” Gardner concentrates on a single creative life, namely, that of Sigmund
Freud. In his analysis of creativity, Gardner distinguishes among four levels of
analysis. The subpersonal includes the biological subtraits and genetic endowments
responsible for creative performance. The personal includes the cognitive processes
and representations responsible for creative performance. It also includes factors of
personality, motivation, and drive. The extrapersonal includes domain-specific fac¬
tors responsible for the emergence of creative performance. Finally, the multiper¬
sonal includes the social context in which individuals make their creative contribu¬
tions. Gardner analyzes Freud’s performance in terms of each of these levels and
helps us understand the multiplicity of factors that, in confluence, produced such a
creative outpouring of work.
In the second section of the third part of this book, the focus shifts to the
employment of historical data from many individuals in order to understand the
antecedents and bases of creative performance.
The first chapter in this section is by Csikszentmihalyi, and it views the interac¬
tions of society, culture, and person to produce a systems view of creativity.
According to Csikszentmihalyi, the creative individual takes information provided
by the culture and transforms it, and if the changes are deemed valuable by society,
they will be included in the domain in which the individual works, thus providing a
new starting point for the next generation of persons. The actions of all three
systems - person, domain (symbol system), and field (social organization of do¬
main) - are necessary for creative performance to occur. Csikszentmihalyi reviews
a variety of creative contributions and shows how his model can be applied to
understand how they took place and how they came to be viewed as creative.
In his chapter “Creativity and Talent as Learning,” Walberg suggests a novel
theory of creativity called “human-capital theory.” According to Walberg,
“human capital” refers to the quality of work or the motivation, skills, and
creativity of the worker. Human capital can be general or specialized. General
capital includes knowledge and skills acquired in early life, mastery of the liberal
arts and sciences, and so on. Specialized capital includes more specific skills and
knowledge acquired only by the few, rather than the many. Using notions such as
6 Introduction

these, Walberg proceeds to an essentially economic analysis of creativity that is

rather different from any of the other analyses in this book. For example, he notes
the relevance of Merton’s “Matthew effect’’ for understanding creativity. The basic
idea behind the Matthew effect is that the rich get richer and the poor get poorer.
The relevance of this notion is that individuals who have shown creative perfor¬
mance are likely to create and encounter the kinds of circumstances that will lead to
further creative performance, whereas those individuals who thus far have not
displayed creative performance are less likely to encounter or create such condi¬
tions. Walberg shows that economic analysis can account for certain aspects of
creative performance that could be accounted for only with great difficulty through
a standard psychological analysis.
The next chapter in this section, by Perkins, discusses “The Possibility of In¬
vention.” It bears a certain resemblance to Johnson-Laird’s chapter in its heavy
emphasis on philosophical analysis, and indeed Perkins is a philosopher by training.
Perkins opens with a consideration of the ex nihilo problem, namely, how some¬
thing can come out of nothing. His main mission in this chapter is to offer some
broad answers to this question. He considers some standard answers to the question
and rejects them. He then proceeds to construct his own answer, which is an
evolutionary one: The main mechanisms of creative invention, according to Per¬
kins, are generation, selection, and preservation. In essence, Perkins proposes that
natural selection can be viewed as applying to the evolution of creative thought as
well as to the evolution of species. At the same time, Perkins recognizes that the
selection process for human creativity is much more strategic than what he refers to
as nature’s rolling of the genetic dice. Perkins devotes considerable attention in his
chapter to the concept of boundaries, precisely because creativity involves going
beyond boundaries. He also describes some of the cognitive heuristic processes
people can use in order to achieve creative performance. Creative solutions rarely
involve considering all possible solutions to a problem, and hence a creative indi¬
vidual needs a set of heuristics for deciding how to avoid combinatorial explosion in
seeking solutions to problems. Perkins describes some of the methods for doing
this, such as hill climbing and parallel processing.
The last chapter in this section, and indeed the last of the main chapters in the
book, is by Simonton. Simonton opens his chapter with two major propositions
concerning creativity. The first is that creativity is a form of leadership in that it
entails personal influence over others, and the second is that creativity involves the
participation of chance processes both in the origination of new ideas and in the
social acceptance of those ideas by others. Simonton builds his chance-configura¬
tion theory around these two propositions. This theory, like that of Perkins, in¬
volves evolutionary thinking. Simonton proposes that solutions to novel problems
require some means of generating ideational variation, that these heterogeneous
variations must be subjected to a consistent selection process that deletes all but
those solutions that feature adaptive fit, and that the variations that have been
selected must be preserved and reproduced by mechanisms of retention. Simonton
Introduction 7

shows mathematically as well as psychologically how these factors combine to

produce, and render effective, creative performance. Simonton also summarizes his
innovative historiometric method, in which historical variables serve as bases for
predicting when and where creativity is more and less likely to take place.
The fourth and last part of the book contains an overview and integration by
Tardif and Sternberg. The goal of this chapter is to organize the main themes in the
theorists’ various views of creativity and to suggest a variety of commonalities that
hold together their various points of view. The commonalities are prominent enough
to suggest that a unified field of creativity may eventually be possible that will
integrate the various perspectives advocated in this volume.
'
Part I

The role of the environment in creativity

1 The conditions of creativity

Beth A. Hennessey and Teresa M. Amabile

What is the origin of creative production? Can it be entirely attributed to a specific

ability or constellation of personality traits possessed by the creator? Or are there
other ingredients equally essential to this process? Our research demonstrates that
social and environmental factors play major roles in creative performance. We have
found that there exists a strong and positive link between a person’s motivational
state - motivational orientation, if you will - and the creativity of the person’s
performance. And in large part it is the social environment, or at least certain
aspects of that environment, that determines this orientation. Guiding our investiga¬
tions is what we have termed the intrinsic motivation principle of creativity.
People will be most creative when they feel motivated primarily by the interest, enjoyment,
satisfaction, and challenge of the work itself - not by external pressures.

In essence, we are saying that the love people feel for their work has a great deal
to do with the creativity of their performances. This proposition is clearly supported
by accounts of the phenomenology of creativity. Most reports from and about
creative individuals are filled with notions of an intense involvement in and un¬
rivaled love for their work. Thomas Mann, for example, described in one of his
letters his passion for writing (John-Steiner, 1985), and physicists who were close
to Albert Einstein saw in him a similar kind of intensity. In the words of the Nobel
Prize-winning inventor Dennis Gabor, “no one has ever enjoyed science as much
as Einstein’’ (John-Steiner, 1985, p. 67).
Similar references to the sustaining delights of creative endeavor are found
throughout the letters, biographies, and autobiographies of a great many creative
individuals (Amabile, 1983; Rosner & Abt, 1970). As cinematic director Henry
Jaglom describes it, the joy of creating is like “being on a bicycle going down hill”
(John-Steiner, 1985, p. 67), and Csikszentmihalyi (1975) describes the “flow
state” experienced by many individuals at the peak of creative insight. Yet this
exhilaration is in no way automatic. In our research we have found that the love felt
for one’s craft can be quite delicate and easily overshadowed by pressures in the
environment.
As early as 1954, Carl Rogers talked about the “conditions for creativity” and
the importance of setting up situations of psychological safety and freedom, of

11
12 B. A. Hennessey and T. M. Amabile

providing an environment in which external evaluation is absent. It was Rogers’s

contention that creativity can flourish only in a climate in which the motivation to
produce comes from within, and an examination of creative individuals’ first-person
accounts of their work supports this viewpoint. Over and over again, reference is
made to the delicate balance needed between the desire for attention, praise, and
support from friends, supervisors, editors, or colleagues, on the one hand, and, on
the other hand, the necessity to maintain a certain protective distance from the
opinions of these very same people (Hughes & McCullough, 1982; Steinbeck,
1969). A concern with evaluation expectation and actual evaluation, a focus on
competition and external reward, a desire for external recognition, the difficulty of
withstanding harsh criticism, a rebellion against time pressures, and a deliberate
rejection of society’s demands - paradoxically, all are recurrent themes in the
phenomenology of outstandingly creative persons.
It is clear from case studies that an environment conducive to creative production
is not easily established, and once achieved, it must be constantly reshaped and
controlled. Love is not enough. The creative individual, often with great difficulty,
must work to construct the personal environment that will offer the fullest oppor¬
tunity for maintaining that love, that intrinsic motivational orientation so necessary
for creative expression (Wallace, 1985, p. 371).

Some definitions

Having presented and briefly discussed the intrinsic motivation principle, we shall
now define our terms a bit more precisely. Intrinsic motivation, which for the
purposes of this discussion will also be referred to as intrinsic interest, has been
variously described by a number of theorists. Harlow (1950) used this term to
describe the interest shown by monkeys in puzzle manipulation. Taylor (1960), in
an “information processing theory of motivation,” suggests an inherent interest in
cognitively engaging in tasks, and a similar construct has been developed by Hunt
(1965).
More recently, some social psychologists (Lepper, Greene, & Nisbett, 1973) have
defined intrinsic motivation from a cognitive perspective: Individuals are intrinsically
motivated if they perceive themselves as engaging in activities primarily because of
their own interest in those activities; they are extrinsically motivated if they perceive
their engagement as a means to achieve extrinsic goals. Kruglanski (1975) expands
on this notion with the addition of an endogenous/exogenous framework. According
to this model, an actor’s inference of intrinsic motivation follows whenever perfor¬
mance of a task appears to be a goal in itself. In this endogenous attribution, the
implication of enjoyment, interest, and so on, derives from the fulfillment of the
actor’s desire. To the extent that this individual attributes the action to something
outside of (exogenous to) the task, however, it implies the individual’s negative affect
- that is, the activity was a means rather than an end, and an exogenous attribution
results.
Conditions of creativity 13

Other theorists conceptualize intrinsic motivation more organismically. White

(1959) and Harter (1978) believe that intrinsic motivation is based on the innate
human need for competence in meeting optimal challenges, and deCharms (1968)
proposes that intrinsically motivated behaviors stem from a desire to experience
personal causation (also referred to as self-determination). In the models proposed by
Deci and Ryan and their colleagues (Deci, 1971; Deci, 1975; Deci & Porac, 1978;
Deci & Ryan, 1985), both a competence component and a self-determination compo¬
nent are included, but self-determination is highlighted as especially essential.
All of these definitions share the same fundamental elements: Persons who engage
in activities because of their own interest or personal sense of satisfaction and
fulfillment are intrinsically motivated, whereas persons who engage in activities to
achieve some goal external to task engagement are extrinsically motivated. The
intrinsic motivation principle on which we base our work contends that it is these
differences in motivation that can lead to significant differences in creative
performance.
The work of other psychological theorists suggests that creativity will most likely
result from an intrinsically motivated state. Crutchfield (1962) proposes that “task-
involved,” intrinsic motivation will lead to higher levels of creativity than “ego-
involved,” extrinsic motivation, and in fact, personality data collected by
Crutchfield (1961) indicate that high levels of intrinsic motivation accompany the
work of notably creative persons. McGraw (1978) also suggests that performance
on heuristic tasks (creativity tasks) should be adversely affected by increases in
extrinsic motivation, whereas performance on algorithmic tasks should be en¬
hanced. In an effort to explain this relationship between motivational state and
creativity of performance, Lepper and Greene (1978) suggest that the intrinsically
motivated person, as opposed to the extrinsically motivated person, should feel
freer to take risks because those risks carry virtually no liability, save any that is
self-imposed.
Finally, Deci and Ryan (1985) state that when people are intrinsically motivated,
they will seek situations that interest them and that require the use of their creativity
and resourcefulness.
But what do we mean by creativity? Psychologists have approached this problem
of definition from a variety of angles and in the past have tended to center their
discussions around either the creative person or the creative process. Although
many contemporary theorists think of creativity as a process and look for evidence
of it in persons (Stein, 1974-1975), their definitions most frequently use charac¬
teristics of the product as the distinguishing signs of creativity. Bruner (1962), for
instance, views the creative product as anything that produces “effective surprise”
in the observer, as well as a “shock of recognition” that the product or response,
though novel, is entirely appropriate.
This combination of novelty and appropriateness is found in most conceptual
definitions of creativity. An emphasis on the product is today widely regarded as the
most useful for creativity research. Most assessment methods, however, use quite
14 B. A. Hennessey and T. M. Amabile

narrowly restricted “products” (Check, 1969; McGraw & McCullers, 1979; San¬
ders, Tedford, & Hardy, 1977). In the majority of creativity studies, performance
on creativity tests is used as the criterion, and most creativity tests are constructed
and scored similarly to tests of verbal and mathematical intelligence.
There does exist one small group of measures that entail the rating of actual
products that were made in response to open-ended instructions (Getzels & Csiks-
zentmihalyi, 1976; Simonton, 1980; Sobel & Rothenberg, 1980). Yet most of these
measures present problems too, when the needs of the social-psychological re¬
searcher are considered. If we are to examine the effects of social-environmental
factors on task motivation and creativity, the influence of domain-relevant skills
must be carefully controlled so that the environmental effects will be detectable
above individual differences in talent or experience. However, most of these assess¬
ment procedures that have judges rate actual single products (such as artworks) have
proved too sensitive to large and stable differences in performance, thus threatening
to swamp the more subtle (but no less important) influence of social environment.
This problem of most assessment techniques in current usage is specific to social-
psychological investigations. There is another large problem, however: Most
creativity investigators, whether relying on creativity tests or on subjective assess¬
ments of products, have conducted their research in the absence of clear operational
definitions.

The consensual assessment technique

Like most creativity researchers, we rely on a product definition: A product is

viewed as creative to the extent that it is both a novel response and an appropriate,
useful, correct, or valuable response to an open-ended task. What differentiates our
approach is that we have gone on to formulate a specific operational definition of
creativity that relies on a consensus of experts: A product or idea is creative to the
extent that expert observers agree it is creative (Amabile, 1982b).
In the application of this operational definition and consensual assessment pro¬
cedure, several important assumptions are made about the nature of creativity and
creativity judgment. First, of course, we assume that products or observable re¬
sponses must be the ultimate hallmarks of creativity and that it is not possible a
priori to specify which objective features of new products will, in fact, be consid¬
ered “creative.” Rather, as is the case with most concepts tied to social-psycholog¬
ical research (Gergen, 1982), criteria for creativity require a historically bound
social context.
Another assumption guiding our operational definition is that although creativity
in a product may be difficult to characterize in terms of specific features, and
although it is difficult to precisely characterize the phenomenology of observers’
labeling of certain products as creative (Feldman, 1980), creativity is something
that people can recognize and often agree on, even when they are given neither a
guiding definition nor a list of specific features to be considered (Barron, 1965). In
Conditions of creativity 15

addition, in accord with previous theorists (Simon, 1967), we propose that there is
one basic form of creativity, one basic quality of products that observers are re¬
sponding to when they label something “creative,” regardless of whether they are
considering works produced in the scientific or artistic domain.
Finally, we assume that there are degrees of creativity, that observers can say
with an acceptable level of agreement that some products are more or less creative
than others. This assumed existence of a continuous underlying dimension is not
new to psychological theorizing on creativity. Cattell and Butcher (1968), for
example, stated that creativity “may be manifested ... at widely differing levels,
from discovering the structure of the atom to laying out a garden” (p. 279). And
Nicholls (1972), while arguing against the assumption of a normally distributed
personality trait of creativity, does concede that the assumption of continuity in
judgments of creative products is a reasonable one.
In our own investigations, this consensual assessment procedure is implemented
as follows: Subjects are asked to complete some task in a specific domain (such as
poetry), and then experts in that domain (such as poets) independently rate the
creativity of the products. The level of interjudge agreement is assessed, and if it is
acceptable (generally above .70), the mean across-judge creativity ratings are used
as our dependent measures of creativity. Before presenting studies that have em¬
ployed this technique, it is necessary to describe the major features of our assess¬
ment method in more detail.
In selecting an appropriate creativity task, there are three requirements that must
be met. First, the task must be one that leads to some product or clearly observable
response that can be made available to appropriate judges for assessment. Second,
the task should be open-ended enough to permit considerable flexibility and novelty
in responses. Third, because it is desirable for social-psychological research that
there not be large individual differences in baseline performances on the task, it
should be one that does not depend heavily on special skills - such as drawing
ability or verbal fluency - that some individuals have undoubtedly developed more
fully than others. Certainly, for tasks that do depend heavily on special skills, it is
possible to reduce extreme interindividual variability by choosing only individuals
with a uniform level of baseline performance. However, this solution is impractical
for most social-psychological studies of creativity. Whereas it is probably advisable
in any case to eliminate subjects with deviantly high or low levels within the domain
being explored, preliminary studies of a laboratory-based nature should use tasks
that virtually all members of the population can perform adequately, without evi¬
dence of large variability in individual differences.
A number of requirements concerning the assessment procedure also deserve
mention. First, judges should all have experience with the domain in question,
although the levels of experience for all judges need not be identical. Basically, the
consensual assessment technique requires that all judges be familiar enough with the
domain to have developed, over a period of time, some implicit criteria for creativ¬
ity, technical goodness, and so on.
16 B. A. Hennessey and T. M. Amabile

The second procedural requirement is that judges make their assessments inde¬
pendently. The essence of the consensual definition is that experts can recognize
creativity when they see it and that they can agree with one another in this assess¬
ment. If experts in a domain say (reliably) that something is highly creative, we
must accept it as such. The integrity of the assessment technique depends on
agreement being reached without attempts by the experimenter to impose particular
criteria or attempts by the judges to influence each other. Thus, the judges should
not be trained by the experimenter to agree with one another, they should not be
given specific criteria for assessing creativity, and they should not have the oppor¬
tunity to confer while making their assessments.
Third, in preliminary work on developing the technique for a given task, judges
should make assessments on other dimensions in addition to creativity. Minimally,
they should make ratings of the technical aspects of the products in question and, if
appropriate, their aesthetic appeal as well. This will then make it possible to deter¬
mine whether, in subjective assessments, creativity is related to or independent of
those dimensions. Assessments of these other aspects of the work also make it
possible to compare social-environmental effects on these aspects with social-en¬
vironmental effects on creativity. This is of particular theoretical importance be¬
cause, theoretically, there might be reasons to predict that a given social factor will
have differential effects on creativity and on technical performance.
Fourth, judges should be instructed to rate the products relative to one another,
rather than rating them against some absolute standard they might hold for work in
their domain. This is important because, for most investigations, the levels of
creativity produced by the “ordinary” subjects who participate will be low in
comparison with the greatest works ever produced in that domain.
Finally, each judge should view the products in a different random order, and
each judge should consider the various dimensions of judgment in a different
random order. If this procedure were not followed and all judgments were made in
the same order by all judges, high levels of interjudge reliability might reflect
method artifacts.
Once the judgments are obtained, ratings on each dimension should be analyzed
for interjudge reliability. In addition, if several subjective dimensions of judgment
have been obtained, these should be factor-analyzed to determine the degree of
independence between creativity and the other dimensions. Finally, if the products
lend themselves to a straightforward identification of specific objective features,
these features may be recorded and correlated with creativity judgments.
As implied by the consensual definition of creativity, the most important criterion
for this assessment procedure is that the ratings be reliable. By definition, interjudge
reliability in this method is equivalent to construct validity. If appropriate judges
independently agree that a given product is highly creative, then it can and must be
accepted as such. In addition, it should be possible to separate subjective judgments
of creativity from judgments of technical goodness, aesthetic appeal, and so forth.
In an effort to firmly establish the utility of the consensual assessment technique,
Conditions of creativity 17

we have carried out a program of research using the technique for assessment of
both verbal and artistic creativity. In over 30 separate investigations, judges have
been asked to assess collages, stories, and poems produced by both children and
adults - all with the same results. Interjudge reliability on ratings is consistently
high, and ratings of creativity generally are separable from ratings of product
technical goodness, clarity, and so forth. (See Amabile, 1982b, for more details on
the use of the consensual assessment technique.)
The hypothesis guiding most of our research is that persons assigned to condi¬
tions of social and/or environmental constraint will perform with an extrinsic moti¬
vational orientation, and their products will, on the average, be significantly less
creative than those of subjects assigned to control (no-constraint) groups. The
sources of evidence relevant to this hypothesis include studies of the effects of
constraint on subjects’ intrinsic motivation and studies of the effects of constraint on
subjects’ creativity and other related qualitative aspects of performance.

Social and environmental influences

Intrinsic motivation

The number of studies investigating the impact of social-environmental constraints

on motivation is substantial, and researchers continue to contribute to our under¬
standing of this subject at a steady rate. The first published studies designed to
consider this issue (Deci, 1971, 1972; Lepper et al., 1973) often involved a one-
session paradigm. Experimental subjects generally worked for 1 hr at a task such as
the Soma or hidden-figures puzzles under some condition of reward, feedback, or
constraint, whereas control subjects performed in the absence of such constraint. In
order to obtain the dependent measure of intrinsic motivation, the experimenter then
left the subject for a period of 8 to 10 min under some credible pretext, such as
needing to use the computer or having to get additional evaluation forms. The
experimenter’s absence created the free-choice period in which subjects were alone
and had no extrinsic reasons for working on the activity. Because subjects were
unaware that they were being observed, and because they had a variety of interest¬
ing things to do, the time they spent with the target activity was used as the measure
of their intrinsic motivation.
Employing this and similar paradigms, researchers have successfully demon¬
strated the undermining effects of a variety of constraints. The impact of monetary
rewards has received the greatest amount of attention, and the evidence is clear. The
experience of performing a task for money significantly decreases subjects’ intrinsic
motivation for that activity (Calder & Staw, 1975; Deci, 1972; Pinder, 1976;
Pritchard, Campbell, & Campbell, 1977). Yet monetary payment is not the only
type of reward that has been observed to have such deleterious effects. Widely
ranging varieties of reward forms have now been tested, with everything from good-
player awards to marshmallows producing the expected decrements in intrinsic
18 B. A. Hennessey and T. M. Amabile

motivation (Greene & Lepper, 1974; Harackiewicz, 1979; Kemoodle-Loveland &

Olley, 1979; Ross, 1975). Finally, a situation devised by Lepper, Sagotsky, Dafoe,
and Greene (1982) was designed to determine whether or not one activity presented
as a means to doing another activity would undermine subsequent intrinsic moti¬
vation, regardless of which activity was the means and which was the end. It was
found that regardless of the specific task used, intrinsic motivation was undermined
for the task that had been presented as the means.
It is important to note that other widely varying external constraints, in addition
to reward, have also been found to decrease intrinsic motivation. Lepper and
Greene (1975), for example, investigated the effects of surveillance on the intrinsic
interest of preschoolers. The knowledge that they were being watched via video
cameras had a significant undermining effect on the children’s intrinsic motivation
for solving puzzles. A 1980 investigation conducted by Thane Pittman and his
colleagues has since extended these findings with the demonstration that direct
surveillance by another individual will also decrease children’s intrinsic motivation
for a play activity.
One of our own early investigations examined the effect of deadlines. In this
study (Amabile, DeJong, & Lepper, 1976), college students given deadlines while
working on interesting word games demonstrated less subsequent intrinsic moti¬
vation in the experimental task than did their control (no-deadline) counterparts.
Detrimental effects of expected evaluation on intrinsic interest have been reported
by Smith (1974). In an expansion of this paradigm, Kamiol and Ross (1977)
examined the effects of both informative and noninformative rewards. They found
that although performance-contingent rewards (rewards delivered only if task per¬
formance meets some predetermined standard) did not significantly increase the
intrinsic motivation of children above the level shown by nonrewarded controls,
such informative rewards did at least maintain motivation relative to a task-
contingent reward group (where rewards were noninformative). Thus, unless re¬
wards contain explicit information about competence, they tend quite consistently
to undermine intrinsic motivation.
These studies and others like them illustrate how extrinsic constraints can under¬
mine subjects’ motivation. The evidence is most convincing, even overwhelming.
(For in-depth reviews of this literature, see Bates, 1979; Condry, 1977; Deci &
Ryan, 1980, 1985; Kruglanski, 1978.) The fundamental question that has guided
our program of research is whether or not these same factors can also affect sub¬
jects’ creativity. Although relatively few other investigators have addressed this
problem directly, there has been research on aspects of performance related to
creativity.

Effects of constraints on creativity-relevant aspects of performance

Each of the studies to be described in this section relied on a modified version of the
standard overjustification paradigm (Deci, 1972), with one important distinction.
Conditions of creativity 19

Here, in addition to intrinsic motivation, qualitative aspects of subject performance

are also assessed. In order to be appropriate for testing hypotheses about the social
psychology of creativity, the experimental task not only must be intrinsically in¬
teresting but also must meet a number of criteria: (1) The task must be an open-
ended (heuristic) task, (2) it must not depend heavily on specialized skills, and (3) it
must be a task in which subjects produce an observable product or response that can
be recorded and later presented to experts to be judged on creativity.
In a study of the effect of reward on children’s artistic creativity, Lepper et al.
(1973) found that for preschoolers who initially displayed a high level of intrinsic
interest in drawing with felt-tip markers, working for an expected “Good Player
Award” decreased their interest in the task. When compared with an unexpected-
reward group and a control (no-reward) group, the subjects who had made their
drawings in order to receive a Good Player Award spent significantly less time
using the markers during free-play periods, and this decrement in interest persisted
for at least a week beyond the initial experimental session. Furthermore, the
globally assessed quality (as rated by teachers) of drawings made by children
expecting a reward was lower than that for the unexpected-reward group or the
control group.
Taking a somewhat similar approach, Greene and Lepper (1974) again offered
rewards to preschoolers under one of two conditions: For one group of children,
receipt of the Good Player Award was contingent only on task performance, where¬
as for the second group the reward was contingent on the “quality” of perfor¬
mance. Once again, children working under the reward conditions produced draw¬
ings of lower quality than their nonrewarded counterparts. The reduction in intrinsic
interest in the drawing task was also replicated.
In a 1979 study, Kemoodle-Loveland and Olley also examined the effect of
expected reward on preschoolers’ drawings. Subjects were first observed during
free-play periods and rated for initial interest in the target activity. During the main
experimental session, one-half of the children were rewarded for their participation,
whereas for the control group no reward expectation was established. Measures of
time spent drawing during free periods and drawing quality were then collected 1
and 7 weeks later. Children with a high level of initial motivation for the drawing
task lost interest when rewarded, whereas the same reward contingency caused
children with initially low levels of motivation to gain interest. Further, at the time
of reward, high-interest subjects drew more products than did their high-interest
nonreward peers, yet their pictures were judged to be of poorer quality. Low-
interest subjects, when rewarded, also drew more than did their low-interest non¬
rewarded counterparts, but for this group the quality of performance was not
affected. This study raises interesting questions about the interaction between
environmental factors and the individual’s baseline level of interest, an issue that we
shall pursue later in this chapter. For the purpose of our present discussion, howev¬
er, it is important to note that decrements in interest were accompanied by decre¬
ments in drawing quality.
20 B. A. Hennessey and T. M. Amabile

Employing a very different experimental task, Garbarino (1975) asked fifth- and
sixth-grade girls to teach a matching task to girls in the first and second grades. The
older children who served as teachers either were promised a reward (a free movie
ticket) or were told nothing of reward. Two raters then observed the tutoring
sessions and made independent assessments across an especially broad range of
qualitative performance dimensions. These dependent variables included the fol¬
lowing: the tutors’ use of evaluation, hints, and demands; the learners’ perfor¬
mances; the emotional tone of the interaction, including instances of laughter be¬
tween the children during a session; the efficiency of the tutoring (learning per unit
of time spent).
Overall, rewarded tutors conducted sessions that were high-pressured and busi¬
nesslike, whereas nonrewarded tutors held sessions that were relaxed and yet highly
efficient. The subjective ratings made by the two observers characterized the re¬
warded sessions as tense and hostile, and the nonrewarded sessions as warm and
relaxed. In addition, the rewarded sessions were marked by more demands from the
tutors, more negative evaluative statements by the tutors, less laughter, and poorer
learning by the younger students.
Finally, Pittman, Emery, and Boggiano (1982) found that nonrewarded subjects
showed a strong subsequent preference for complex versions of a game, whereas
rewarded subjects chose simpler versions. And Shapira (1976) reported that sub¬
jects expecting payment for success chose to work on relatively easy puzzles,
whereas subjects expecting no payment preferred much more challenging puzzles.
Each of these investigations points to the same conclusions as do the original
findings of Lepper et al. (1973). For subjects who initially display a high level of
interest in a task, working for an expected reward decreases their motivation and
undermines the globally assessed quality of their performance.

Negative effects of constraints on creativity

More recently, several studies have focused specifically on creative aspects of

subjects’ performances. One of the earliest investigations of this type was conceived
by Kruglanski, Friedman, and Zeevi (1971). Israeli high school students who either
had or had not been promised a reward (a tour of the Tel Aviv University Psychol¬
ogy Department) were given two open-ended creativity tasks. These tasks, adapted
from Barron (1968), required subjects to list as many titles as possible for a literary
paragraph and to use as many words as possible from a 50-word list in writing their
own story.
Product originality was assessed by two independent judges, with good inter¬
judge reliability, and a clear and statistically significant superiority for nonrewarded
subjects emerged. In addition, nearly significant differences were found between
the two groups on two intrinsic measures: subjects’ expressed enjoyment of the
activities and their willingness to volunteer for further participation.
Like the majority of studies that have examined the effect of reward on intrinsic
Conditions of creativity 21

motivation, these investigators used a paradigm whereby subjects were promised

some tangible reward before engaging in an activity, but the reward itself was not
delivered until after that activity had been completed. Critics of overjustification
research (Reiss & Sushinsky, 1975) have suggested that this procedure allows for
alternative explanations of subsequent declines in task engagement. They suggest,
for example, that subjects are so distracted by the anticipation of reward during the
initial task engagement that their enjoyment of the activity is hampered. Thus,
according to this explanation, later lack of interest in the activity occurs not because
subjects have come to see it only as a means to some external goal but because their
intrinsic enjoyment of the activity is directly blocked by the “competing response”
of reward anticipation.
In one of our investigations of the effect of reward (Amabile, Hennessey, &
Grossman, 1986, Study 1), we used a paradigm that renders the competing-re-
sponse hypothesis untenable. The reward offered to children before task engage¬
ment was not a tangible gift to be delivered afterward. Instead, it was an attractive
activity - playing with a Polaroid camera - that the children were allowed to engage
in before completing the target task. In other words, children assigned to the reward
condition promised to do the target activity in order to first have a chance to use the
camera. Children in the no-reward condition were simply allowed to use the camera
and then were presented with the target task; there was no contingency established
between the two. Thus, because children in both conditions had already enjoyed the
“reward” before the experimental activity began, it is unlikely that any reward-
related competing responses were operating.
In a broad attempt to examine the mechanism behind this undermining of
creativity, we also included a number of other manipulations and measures in our
design. Subjects’ levels of intrinsic interest were assessed through self-report and
behavioral observation during a free-play period. And in addition to the offer of
reward, task labeling (as work or play) was introduced as a second dependent
variable. This was done in order to test the possibility that an affective response
such as simply viewing activities as “work” can lead to the same undermining
effects on intrinsic motivation and creativity as being placed under extrinsic
constraints.
Finally, we tried to test the proposition that the undermining effect arises from a
discounting of intrinsic interest when salient extrinsic forces are present (Bern,
1972; Lepper et al., 1973). To do this, we assessed the discounting capabilities of
each of the children and set out to determine whether or not the reward contingency
had different effects on subjects not yet capable of this cognitive process.
In order to examine the impact of reward expectation on children’s verbal
creativity, the elementary school subjects in this study were asked to tell a story to
accompany a set of illustrations in a book with no words. They did this by saying
“one thing” about each page into a tape recorder. Like all the creativity tasks used in
our research, this storytelling activity was specifically designed with two goals
in mind. First, it was necessary that the importance of individual differences in
22 B. A. Hennessey and T. M. Amabile

domain-relevant skills be minimized, because these could lead to high variability

in baseline performances. In the storytelling task, for example, differences in chil¬
dren’s verbal fluency were minimized by restricting their responses to one sentence
per page. Second, in order to be appropriate for testing hypotheses about creativity,
the task had to allow for a wide variety of responses. In other words, the target activity
had to be an open-ended activity (Amabile, 1982b; McGraw, 1978).
At the beginning of the experimental session, children in grades 1 through 5 were
given the opportunity to take two pictures with an “instant” camera. In order to
have a chance to use the camera first, subjects in the reward condition promised to
tell the story later. So that this contingency would be especially salient, children in
this condition were asked to write their names on a piece of paper, a contract also
signed by the experimenter. Subjects in the no-reward condition simply took the
pictures and then told the story; there was no contingency established between the
two tasks. In an application of the consensual creativity assessment technique
(Amabile, 1982b), elementary school teachers familiar with children’s writing later
rated the stories for creativity, with a high level of interjudge reliability. Results
indicated that, overall, children in the no-reward conditions told more creative
stories than did children in the reward conditions. This main effect of reward was,
in fact, statistically significant.
The results indicated that the work-play task labels had no effect on children’s
creativity or self-reported affect, nor did we find any differential effects of the
reward manipulation on the creativity or intrinsic interest of subjects not yet able to
discount. This study does, however, contribute to our understanding of the under¬
mining effect of reward in one important respect. It is significant that this undermin¬
ing effect occurs even when nonrewarded subjects also experience the “reward”
and even when the reward is delivered before the target activity. The only difference
in the experiences of the rewarded and nonrewarded children in this paradigm was
their perception of the reward as contingent or not contingent on the target activity.
It appears that the perception of a task as the means to an end is the crucial element
for creativity decrements in task engagement.
Another of our investigations (Amabile, 1982a) also examined the effect of
expected reward, but in this case the experimental task involved artistic creativity,
and the reward was introduced in a competitive setting (incorporating an evaluative
element). Girls whose ages ranged from 7 to 11 years made paper collages during
one of two parties held in the common room of their apartment complex. Subjects in
the experimental group competed for prizes, whereas those in the control group
expected that the prizes would be raffled off.
Artist-judges later rated each collage on creativity, with a high interjudge reliabil¬
ity. The control group was judged significantly higher than the experimental group
on creativity of collages. Once again, these results are consistent with the proposi¬
tion that intrinsic motivation is conducive to creativity, whereas extrinsic moti¬
vation is detrimental.
Conditions of creativity 23

Several intrinsic motivation theoriests have proposed that in order to undermine

intrinsic interest, a reward must be salient (Ross, 1975) and must be perceived as a
means to an extrinsic end (Calder & Staw, 1975; Deci, 1975; Kruglanski et al.,
1971; Lepper et al., 1973). One method of demonstrating the crucial role of per¬
ceiving a task as a means to an end is to offer subjects a choice concerning task
engagement. Although choice should influence motivation in a manner similar to
that for other social-environmental variables, there are important differences be¬
tween choice and other factors such as expectation of reward or evaluation. If
subjects perceive themselves as freely choosing to perform an activity for which a
reward is offered (or an evaluation promised), they may well adopt an extrinsic
motivational orientation toward that activity - viewing the task as work, and their
own engagement as motivated by external pressures. On the other hand, if subjects
are simply presented with a task and told that they will be paid or evaluated, with no
choice in the matter, we should not expect to observe this same detrimental effect.
This phenomenon has in fact been demonstrated in a study conducted by Folger,
Rosenfield, and Hays (1978).
Some of our investigations have also examined the role of choice. In a prelimi¬
nary test of the effect of choice on creativity (Amabile & Gitomer, 1984), nursery
school children were asked to make paper collages. Children assigned to the choice
condition were allowed to choose any 5 out of 10 boxes of materials to use in this
task. For children in the no-choice condition, however, the experimenter made the
box selections - yoking the two conditions in a matched-pairs design. All subjects
then completed their collages, which were later rated for creativity by artists.
Two weeks after this initial session had ended, a behavioral measure of subse¬
quent intrinsic interest in the collage activity was obtained. Over a 3-day span,
leftover materials were made available, and each child’s engagement in the collage
activity during free-choice periods was timed. As predicted, there was a substantial
difference in collage creativity between the conditions. Subjects in the choice condi¬
tion made collages that were significantly more creative than those made by subjects
in the no-choice condition. In addition, children in the choice condition spent
somewhat more time with collage materials during free play than did children in the
no-choice condition.
In an investigation of the interactive effects of reward and choice on children’s
verbal creativity (Amabile et al., 1986, Study 2), the meaning of “reward” was
directly manipulated by varying its method of presentation. We employed a 2 x 2
factorial design in which the presence or absence of choice was completely crossed
with the presence or absence of reward. We predicted that creativity would be
undermined only in those children who explicitly contracted with the experimenter
to make the collage in order to obtain the reward.
Subjects in this investigation were 80 students in grades 3,4, and 5. All children
participated in the experimental session individually. Subjects in the choice-reward
condition were told that either they could go back to their classrooms or, if they
24 B. A. Hennessey and T. M. Amabile

signed a contract promising to make a collage and tell a story, they could first take
two pictures with an instant camera as a reward. Importantly, it was made clear that
some children did in fact decide to leave.
In the choice-no-reward condition, the children were told that they could do the
things described or go back to their classroom, and again it was made clear that the
decision to go back to class was an acceptable response. In the no-choice-reward
condition, subjects were told that because they were going to do things for the
experimenter, they would first be allowed to take two pictures with the camera. And
in the no-choice-no-reward condition, the picture taking was simply introduced as
one in a series of tasks to be completed. The two tasks that followed, collage
making and storytelling, were presented in different counterbalanced orders.
After they had completed each of these activities, all children were asked to
answer two questions. Using a continuum of circles representing a range of re¬
sponses from “I liked it a lot” to “I didn’t like it at all,” subjects each marked the
circle that best described how they felt about making the design or telling the story.
They were also asked, “If you were to tell another kid about what you did here,
would you say you worked or you played?”
A 2 x 2 ANOVA on the creativity scores for both the stories and the collages
revealed the predicted interaction between reward and choice. Subjects in the
choice-reward condition produced stories lower in creativity than did subjects in
the other three groups. The creativity of children who had been given a choice about
their participation had been seriously undermined by the reward manipulation,
whereas for the children who did not perceive that the decision to complete the task
was under their control, no such deleterious effects were observed. (Analyses of the
self-report measures of affect and the work-play distinction revealed no significant
effects.) Thus, this study demonstrated that it is not reward per se but rather the
functional significance of reward as controlling the performance that undermines
creativity.
In an effort to conceptually replicate these findings, a second investigation
(Amabile et al., 1986, Study 2) was carried out in which reward and choice were
again crossed in a 2 x 2 factorial design. This study differed from its predecessor in
a number of respects, however. The most important difference was that the subjects
were adult women, rather than young boys and girls. In addition, a monetary reward
was employed, and the money (though shown to rewarded subjects before task
engagement) was awarded only after the task had been completed. The creativity
task used here was the same collage-making activity as that presented to the children
in the original investigation.
The results here provide a close replication of the results in Study 2. A 2 x 2
ANOVA again revealed a significant reward x choice interaction, resulting largely
from the low creativity of subjects in the contracted-for reward group (choice-
reward). As predicted, and as in Study 2, the lowest level of creativity was found in
this condition.
Conditions of creativity 25

None of the items on a post-task questionnaire designed to assess subjects’

interest in and feelings toward the collage activity revealed significant main effects
or interactions of the independent variables. However, correlational analyses do
suggest a relation among intrinsic interest in the activity, enjoyment, and creativity.
Having established the deleterious effects of reward and conditions of restricted
choice, what other extrinsic constraints might also have a negative impact on
subjects’ creativity? Our research has consistently demonstrated that the expectation
of evaluation produces similar decrements in subjects’ interest and task perfor¬
mance. The initial study of evaluation expectation (Amabile, 1979) was designed to
test the hypothesis that such constraint will undermine creative performances on
heuristic tasks. The task we used was a collage-making activity, and it was pre¬
dicted that subjects placed under constraint (evaluation expectation) would show
lower levels of both creativity and intrinsic interest in the task than would no¬
constraint controls. The subjects in this study were 95 women enrolled in an intro¬
ductory psychology course at Stanford University. As a cover story, the experi¬
menter explained that this investigation was actually a pretest for another experi¬
ment to be done the following quarter. The alleged purpose of this pretest was to
identify activities that would affect most individuals’ moods in predictable ways.
Subjects assigned to the control condition were told also that their designs would not
be used as sources of data - that the only information of interest was the mood they
reported on a questionnaire. In contrast, subjects in the evaluation condition were
told that their finished designs would be rated by graduate artists from the Stanford
Art Department and would serve as important sources of information.
In introducing the art task, the experimenter stressed that subjects had complete
freedom in using the materials to form a design, but that only the materials provided
should be used. In addition, each subject was asked to make a design that conveyed
a feeling of “silliness,” as when a child is acting and feeling silly. In order to
further reduce extraneous sources of variability, subjects with significant levels of
experience in artistic or collage work were eliminated.
Each subject was left alone for 15 min to work on her collage. At the end of the
session, the experimenter presented the subject with a Mood Questionnaire (in
keeping with the cover story) and an Art Activity Questionnaire that included a
number of questions designed to assess subjects’ interest in and attitude toward the
collage task. The 95 collages were rated by 15 artist-judges following the same
consensual assessment techniques used in our other studies (Amabile, 1982b). The
results supported the hypothesis that evaluation expectation is detrimental to
creativity, with control (nonevaluation) subjects producing significantly more cre¬
ative products than experimental (evaluation) subjects. In addition, control subjects
scored higher in self-rated interest than did experimental subjects.
In order to firmly establish this finding that an expected evaluation has negative
effects on an adult’s creativity, two further investigations were undertaken.
In both studies, a replication with an artistic activity (Amabile, Golfarb, &
26 B. A. Hennessey and T. M. Amabile

Brackfield, 1982, Study 2) and a replication with a verbal activity (Amabile et al.,
1982, Study 1), there was a significant main effect of evaluation expectation on the
creativity ratings. Nonevaluation subjects made collages and wrote poems that were
judged significantly more creative than those produced by evaluation subjects.
In a study done with a group of colleagues (Berglas, Amabile, & Handel, 1981),
we set out to examine the effects of prior evaluation on children’s subsequent
creativity. We predicted that highly salient evaluation on one task would lead
children to expect evaluation on a later task with the same experimenter and, as a
consequence, would lower their creativity on the later task. All subjects, boys and
girls in grades 2-6, made two artworks. The first involved painting with a spinning
disk, and the second (which was the target task) involved making a paper collage.
Experimental-group children were positively evaluated by the experimenter on their
“spin art” before they did the collage. Control-group subjects simply made the two
artworks, with no evaluation. Creativity results indicated a clear superiority of the
control group over the experimental group. In other words, we observed an overall
negative impact of prior evaluation on creativity of performance - even though that
evaluation had been positive.
Finally, Koestner, Ryan, Bemieri, and Holt (1984) set out to determine if the
functional significance of another extrinsic constraint - behavioral limits - affects
children’s motivation and creativity. In this study, 5- and 7-year-old children were
asked to engage in an intrinsically interesting painting activity under three limit¬
setting conditions that varied along an information/control dimension. In the con¬
trolling-limits group, restrictions pertaining to task neatness were stated in terms of
“should” and “must.” The informational group was presented with a verbal
communication conveying the same behavioral constraints in the absence of exter¬
nal pressure, but with an acknowledgment of possible conflicting feelings about the
imposed limits. For the control (no-limits) group, no mention was made of these
constraints.
Following the period during which the children made their paintings, they were
left alone for a free-play session. The amount of time spent during this period was
used as a measure of intrinsic motivation. Finally, the children were asked to rate
how much they enjoyed the painting activity. Paintings made during the main
session were assessed by artist-judges for creativity, using our (Amabile, 1982a)
consensual assessment technique. Results indicated a main effect for condition.
This effect supports the main experimental prediction, in that subjects in the no¬
limits and informational-limits groups spent more free-choice time painting than did
controlling-limits subjects. Thus, intrinsic motivation was significantly greater for
children in the nolimitational and informational conditions than for children in the
controlling condition. In addition, a marginal effect for limit-setting style emerged
on the self-report measure of enjoyment. Effects for limit-setting style were also
found for creativity, with the no-limits group significantly more creative than the
controlling-limits group. The informational-limits group was intermediate.
Conditions of creativity 27

A more positive approach: maintaining creativity by maintaining

intrinsic motivation

In our attempt to demonstrate a definitive link between creativity and intrinsic

motivation, it is just as important to demonstrate that creativity will be maintained
when intrinsic motivation is maintained as it is to demonstrate that creativity will be
undermined when intrinsic motivation is undermined. In a recent study with Barba¬
ra Grossman (in preparation), we set out to determine if special training sessions
designed to directly address motivational orientation could “immunize” children
against the usually deleterious effects of reward on intrinsic motivation and creativ¬
ity of performance. In the first phase of this experiment, students in grades 3,4, and
5 were randomly assigned to one of two conditions: intrinsic motivation training or
control, both of which were run by the same female experimenter.
In the crucial intrinsic motivation training condition, subjects were shown vid¬
eotapes depicting two attractive 11-year-old children talking with an adult about
various aspects of their schoolwork. The scripts for these tapes had been specifically
designed so that the boy and girl on the tape would serve as models of highly
intrinsically motivated individuals. There were two primary messages conveyed by
these intrinsic motivation training tapes. Our first goal was to get the children to
focus on intrinsic reasons for doing work in school and to concentrate on those
aspects for maximal enjoyment. The following is an example of a tape segment that
addresses this issue:

Adult: Tommy, of all the things your teacher gives you to do in school, think about the one
thing you like to do best and tell me about it.
Tommy: Well, I like social studies the best. I like learning about how other people live in
different parts of the world. It’s also fun because you get to do lots of projects and reports. I
like doing projects because you can learn a lot about something on your own. I work hard on
my projects, and when I come up with good ideas, I feel good. When you are working on
something that you thought of, and that’s interesting to you, it’s more fun to do.
Adult: So, one of the reasons you like social studies so much is because you get to learn about
things on your own. And it makes you feel good when you do things for yourself; it makes it
more interesting. That’s great!

The second issue addressed in the intrinsic motivation training tapes was the
practice of cognitively distancing oneself from socially imposed extrinsic con¬
straints - focusing instead on the inherently enjoyable aspects of a task in an effort
to maintain intrinsic motivation in the face of such factors as reward or evaluation.
An example:

Adult: It sounds like both of you do the work in school because you like it, but what about
getting good grades from your teacher or presents from your parents for doing well. Do you
think about those things?
Tommy: Well, I like to get good grades, and when I bring home a good report card, my
parents always give me money. But that’s not what’s really important. I like to leam a lot.
28 B. A. Hennessey and T. M. Amabile

There are a lot of things that interest me, and I want to learn about them, so I work hard
because I enjoy it.
Sarah: Sometimes when I know my teacher is going to give me a grade on something I am
doing, I think about that. But then I remember that it’s more important that I like what I’m
doing, that I really enjoy it, and then I don’t think about grades as much.
Adult: That’s good. Both of you like to get good grades, but you both know that what’s really
important is how you feel about your work, and that you enjoy what you are doing.

In small groups of 3 to 5 members, subjects met with the experimenter for two
20-min training periods on 2 consecutive days. Each intrinsic motivation training
session consisted of showing segments of the videotape, interspersed with directed
discussion. During these discussions, the children were asked to relate what they
had seen on the tape, to answer for themselves the questions the adult had posed,
and to give their own reactions to the content of Tommy’s and Sarah’s responses.
Throughout, the experimenter offered interpretations of the tape and of the chil¬
dren’s commentary and shared her own ideas, all with the aim of making them more
aware of intrinsic motivation and methods of coping with extrinsic constraints. At
the close of each of these brief meetings, the children were asked to complete a
series of short exercises in which they indicated their preferences for a variety of
school activities and described their feelings when performing their favorite tasks.
Subjects assigned to the control group also met in small groups over a 2-day
period for the purpose of viewing videotapes. In this case, however, the discussion
centered around their favorite things: foods, movies, animals, and so forth. In
summary, then, all subjects participated in some form of group activity. All met
with the experimenter, saw videotapes, and participated in group discussions. What
differentiated the conditions was the focus of these sessions: intrinsic motivation or
issues irrelevant to intrinsic motivation.
In the second phase of this experiment, after the training sessions had been
completed, each child met individually with a different experimenter for testing.
(The children’s teachers and the experimenters were careful to avoid mentioning
any connection between the training and testing sessions, and they denied a connec¬
tion if any of the children inquired.) The Harter Scale of Intrinsic versus Extrinsic
Orientation in the Classroom (Harter, 1981) was administered, and two dimensions
of classroom motivation were assessed. These two dimensions, each having an
intrinsic pole and an extrinsic pole, were (1) Curiosity/Interest versus Pleasing the
Teacher/Getting Good Grades and (2) Independent Mastery versus Dependence on
the Teacher.
After this administration, a reward manipulation was introduced. Following a
procedure used in an earlier study (Amabile et al., 1986, Study 1), half of the
children in each of the three training conditions were told that they could take two
pictures with an instant camera if they promised that later they would tell a story for
the experimenter. For the remaining children, this picture taking was presented
simply as the first in a series of “things to do.”
The major dependent measure - creativity on a storytelling activity - also paral-
Conditions of creativity 29

leled that employed in a previous investigation (Amabile et al., 1986, Study 1). As
a final task, the Unusual Uses Test of the Torrance Tests of Creative Thinking
(Torrance, 1966) was administered.
We predicted that for subjects who had been trained to deal effectively with
extrinsic constraints and to focus on intrinsic reasons for doing work in school,
overall intrinsic motivation would be increased. Significant differences were, in
fact, found between the scores on the Harter Curiosity scale for children in the two
treatment conditions. Children receiving intrinsic motivation training scored higher
than subjects in the control condition.
In addition, an examination of story creativity revealed the predicted interaction
between training condition and reward manipulation. Although no main effects
were found, expectation of reward did produce the predicted decrement in creativity
for subjects assigned to the control group. By contrast, students in the intrinsic
motivation condition who were rewarded for their participation told stories that
were judged significantly more creative than those told by the no-reward intrinsic
motivation subjects. It would seem that as a result of their training, these children
had learned to treat reward not as an element that detracts from intrinsic interest but
as something that can add to overall motivation. They had learned to overcome the
deleterious effects of reward - so much so that their levels of intrinsic motivation
(and therefore their levels of creativity) seem to have increased.

Mechanisms for the conditions of creativity

The results of our training study are at the same time extremely exciting and
somewhat puzzling. How can we explain the fact that those children who had
received intrinsic motivation training exhibited higher creativity when rewarded
than when not rewarded? Perhaps the answer lies in their interpretation of the
reward manipulation and storytelling activity. Perhaps our intrinsic motivation
training sessions had caused these young subjects to perceive their situation differ¬
ently in some crucial way than did the control group.
This possibility is, in fact, not without both theoretical and empirical support. In
their recent book Intrinsic Motivation and Self-determination in Human Behavior
(1985), Deci and Ryan observe that previous research has tended to focus primarily
on the outward experimental events themselves: the presence or absence of sur¬
veillance and the nature of the reward structure, for example, and their average
effects on people’s motivation and related variables such as creativity (Deci &
Ryan, 1985, p. 87). It is the belief of Deci and Ryan, however, that the impact of an
event on motivational processes is determined not by the objective characteristics of
the event but rather by “its psychological meaning for the individual” (p. 85).
According to their cognitive evaluation theory, all external events can be viewed
as informational, controlling, or amotivating. An environmental event that is per¬
ceived as controlling is one that is interpreted by the perceiver as pressure to attain a
given behavioral outcome - pressure that is interpreted to induce or coerce the
30 B. A. Hennessey and T. M. Amabile

recipient to perform in a specific manner. When this aspect is salient, the perception
of an external locus of causality is facilitated, and intrinsic motivation tends to be
undermined. An environmental event that is perceived as informational is one that
provides the recipient with behaviorally relevant information in the absence of
pressure to attain a particular outcome. A salient informational event increases
intrinsic motivation if it signifies competence and decreases intrinsic motivation if it
signifies incompetence. Finally, according to cognitive evaluation theory, it is also
sometimes possible to classify some events as internally amotivating. These would
be events occurring within a person, such as self-deprecation or hopelessness, that
signify one’s inability to master certain situations. Whether an event will be per¬
ceived as informational, controlling, or amotivating, Deci and Ryan believe, is an
issue of the relative salience of these three aspects for the perceiver and is affected
by one’s sensitivities and past experiences as well as by the actual configuration of
the event itself (p. 85).
How might this analysis be applied to the specific case of our intrinsic motivation
training group? The message conveyed by our videotapes and guided discussions
was that external rewards such as receiving good grades or money from parents are
nice, but what is really important is that one truly enjoy what one is doing. In
essence, what we had attempted, and evidently accomplished, was to develop a
salient intrinsic orientation, or a more solidly internal locus of control, in our
subjects. Thus, whereas the nontrained subjects perceived the reward manipulation
as strongly controlling, the trained subjects most likely did not.
In our study, we attempted to create individual differences in motivational orien¬
tation between subjects. Other researchers have taken the route of examining dif¬
ferences that already exist. In one of the few investigations that have examined
differences in the perceiver as possible mediators of the effects of extrinsic con¬
straints on intrinsic motivation, Lonky and Reihman (1980) studied the impact of
verbal praise on the intrinsic motivation of children scoring high and low on internal
locus of control. They found that when children high on internal locus of control
were praised, they showed an increase in intrinsic motivation over pretreatment
assessments; but children scoring at the low end of the internal locus scale showed
decreases in intrinsic motivation levels after being praised. These authors conclude
that persons high on internal locus of control believe themselves to be more in
control of outcomes and are more likely to interpret rewards and communications as
informational, whereas persons low on internal control will be more likely to
interpret these same elements as controlling.
Another study with a focus on individual differences was carried out by Boggiano
and Barrett (1984). In this investigation, the effects of positive and negative feed¬
back on the intrinsic motivation and performance of children who differed in their
initial motivational orientation were assessed. It was found that success feedback
increased the intrinsic motivation of intrinsically oriented children, but not that of
extrinsically oriented children. Negative feedback also was observed to signifi¬
cantly increase the motivation of the intrinsically oriented subjects. For this group,
Conditions of creativity 31

it was apparently viewed as a challenge. However, for the extrinsically oriented

children, the same negative message seemed to represent evidence of their incompe¬
tence, decreasing their intrinsic motivation even further and creating feelings of
amotivation and helplessness.
In summary, the research outlined here demonstrates that characteristics of the
perceiver (or recipient) of an environmental constraint, such as reward, have a great
deal to do with how that constraint is received and interpreted. As Deci and Ryan
point out, most situations are ambiguous enough that the relative salience of infor¬
mational, controlling, and amotivational aspects of a situation can be unique for
each individual. The same event can be perceived very differently, depending on
the orientations of the persons involved. It would seem that in our training study,
the intrinsic motivation training sessions sufficiently altered the children’s percep¬
tions of the reward manipulation so that they counteracted the usual undermining
effects on motivation and creativity. Rather than passively receiving the parameters
of the reward situation, it appears that these children actively constructed them.
They seem to have tailored the environment to meet their needs. (Interestingly, this
same ability to shape one’s environment is something that Sternberg, 1985, has
linked to gifted individuals.)
What we are really addressing here are questions that go far beyond the specific
case of our intrinsic motivation training sessions to the broader issue of individual
differences in general. For us, as well as for others, this is a relatively new area for
theorizing and research on the linkage among environmental conditions, moti¬
vation, and creativity. Up until this point, we have been concerned primarily with
global effects - with the general effects of extrinsic constraint on motivation and
creativity across a variety of situations and subject populations. Ours has been a
hydraulic model, not unlike that used by most early intrinsic motivation theorists
(Lepper & Greene, 1978): Intrinsic motivation is conducive to creativity, and ex¬
trinsic motivation is detrimental; as extrinsic constraints increase, intrinsic moti¬
vation and creativity must decrease. Yet, as our research and that of others has
begun to demonstrate, perhaps this formula does not accurately describe all situa¬
tions. Perhaps, under certain circumstances or with certain individuals, intrinsic and
extrinsic forces can combine in an additive fashion. Certainly this appears to be
what happened in the training study. Rather than detracting from the children’s
performances, the offer of reward augmented the curiosity and the creativity of
performance of the intrinsic motivation group. Apart from the internal-external
distinction in locus of control and the related intrinsic-extrinsic orientation distinc¬
tion, might there be other naturally occurring individual differences that would
make some people less vulnerable to the negative effects of extrinsic constraint on
their intrinsic motivation and creativity?
Our observations of and interviews with people in the workplace lead us to
believe that this is so. In a recent interview study (Amabile & Gryskiewicz, in
press), we discovered that research and development scientists working within the
same laboratory sometimes perceive the same extrinsic constraints quite differently.
32 B. A. Hennessey and T. M. Amabile

There are two interesting subgroups in the sample, exemplifying opposite extremes
around the usual “modal” response to constraints: those who feel constantly sup¬
pressed by the constraints in their environment, and those who have somehow
managed to rise above these constraints (or at least manage to view them in a
perspective that does not interfere with creative production). Although we do not
have detailed individual difference measures on these scientists, we can speculate
on what might be the crucial distinguishing characteristics between these two types
of workers.
One dimension that may be relevant is that of self-esteem. In fact, an examination
of the literature reveals that this personality construct may play a significant role
where intrinsic motivation and creativity are concerned. A study conducted by
Deci, Nezlek, and Sheinman (1981) revealed, for example, that children in public
school classrooms run by teachers who were oriented toward supporting autonomy
had higher self-esteem and more intrinsic motivation than children assigned to
classrooms where teachers were oriented toward controlling behavior. Similar find¬
ings were reported by Harter (1982). Ryan and Grolnick (1984) also found strong
positive correlations between intrinsic motivation and self-esteem in children, and
Deci and Ryan (1985) presented data indicating that “strong and stable self-esteem
seems to emanate from a strong sense of self, which motivationally means intrinsic
motivation and more integrated internalization of extrinsic motivation” (p. 142).
The more internalized one’s extrinsic motivation, the more likely it is to contribute
to a sense of positive self-esteem.
The evidence does not stop here. As early as the 1950s, researchers were pointing
to the tendency of creative individuals to display strong self-acceptance and positive
self-evaluation behavior (Fromm, 1959; Guilford, 1950). In a detailed treatise on
the antecedents of self-esteem, Coopersmith (1967) observes:

There thus appears to be an underlying similarity in the processes involved in creative

innovation and social independence, with common traits and postures required for expression
of both behaviors. The difference is one of product - literary, musical, artistic, theoretical
products on the one hand, opinions on the other - rather than one of process. In both
instances the individual must believe that his perceptions are meaningful and valid and be
willing to rely upon his own interpretations. He must trust himself sufficiently that even when
persons express opinions counter to his own he can proceed on the basis of his own percep¬
tions and convictions, (p. 58)

The importance of self-esteem for creative expression appears to be almost beyond disproof.
Without a high regard for himself the individual who is working in the frontiers of his field
cannot trust himself to discriminate between the trivial and the significant. Without trust in
his own powers the person seeking improved solutions or alternative theories has no basis for
distinguishing the significant and profound innovation from one that is merely different. . . .
An essential component of the creative process, whether it be analysis, synthesis, or the
development of a new perspective or more comprehensive theory, is the conviction that one’s
judgment in interpreting the events is to be trusted, (p. 59)

Despite this conviction that self-esteem is an essential prerequisite for creative

expression, Coopersmith could find no studies that had directly investigated this
Conditions of creativity 33

relationship. Recognizing this gap in the literature, he set out to test this hypothesis
and administered three tests to a group of adolescents: Unusual Uses, Circles, and
Draw a Person (Torrance, 1966). His results were especially revealing. Groups high
in subjective self-esteem performed in the most creative fashion on all three bat¬
teries, whereas groups low in self-esteem were significantly less original and inno¬
vating. Coopersmith observed these differences across the variety of conceptual,
linguistic, and artistic skills required in the several tasks and suggested that this
consistency indicates that persons high in self-esteem are likely to be more assert¬
ive, independent, and creative than persons with lower self-esteem - the conclusion
being that individuals with high self-esteem listen to themselves more and are far
more likely to trust their own judgments and reactions (Coopersmith, 1967).
In a similar investigation conducted by Garwood (1964), these predicted rela¬
tionships between personality factors and creativity were again examined, this time
within a population of young scientists. Using the Self-Acceptance scale of the
California Psychological Inventory (CPI) as an index of self-esteem, creative sub¬
jects were observed, as a group, to score significantly higher than their not-so-
creative colleagues.
Although both of these investigations are highly suggestive, they offer only
observational evidence of the connection between creative expression and high
levels of self-esteem. Equally essential to the support of this relationship are data of
a more experimental nature. If it could be shown, for example, that interventions
designed to increase subjects’ creativity can also have a positive effect on their self¬
esteem, or that conditions affecting self-esteem also affect creativity, our case
would be strengthened considerably. Three recent studies have accomplished this
end.
The first (Stasinos, 1984) employed a pretest-posttest control-group design. As a
pretest, the verbal and figural subtests from the Torrance Tests of Creative Thinking
(TTCT) (1968 revised) and two self-esteem instruments designed by Coopersmith
(1967) were first administered to a group of 90 middle and upper school mentally
handicapped Greek children. Subjects assigned to the treatment condition then
received 16 weeks of exposure to the Mark 1, New Directions in Creativity (NDC)
program (Renzulli, 1973). Subjects assigned to the control-group classes continued
regular activities during this time. At the end of this training period, the same
examinations and instruments used in pretesting were again employed. Analyses of
variance yielded the predicted results. As a result of their training, the experimental
group scored significantly higher than did the control group on five of the seven
indices of creativity employed. In addition, the two groups differed greatly in terms
of subjective self-esteem measures, with the experimental group attaining the higher
score, probably as a result of the creativity training program.
Taking a very different approach, Brockner and Hulton (1978) also present strong
evidence of a substantial link between creativity and high levels of self-esteem.
Recognizing a suggestion in the literature that persons low in self-esteem (low SEs)
are more self-conscious than persons high in self-esteem (high SEs) (Ickes, Wick-
34 B. A. Hennessey and T. M. Amabile

lund, & Ferris, 1973; Turner, Scheier, Carver, & Ickes, 1978), they went on to
predict that it is self-consciousness that can impair task performance. It was rea¬
soned that if low SEs could be led to focus their attention away from themselves and
onto the task, performance would improve relative to high SEs. Subjects high and
low in chronic self-esteem performed a concept formation task under three condi¬
tions: (1) in the presence of an audience, where self-focused attention is presumed
to be high; (2) in a control group, in which attention was not manipulated; and (3)
with instructions to concentrate diligently on the task itself. In this, a 2 x 3
between-subjects factorial design, a significant effect was in fact obtained. Low SEs
performed worse than high SEs in the audience condition, no differently in the
control condition, and better than the high SEs when instructed to concentrate on the
task. Brockner and Hulton conclude that the attentional state of low SEs makes
them more susceptible or prone to be adversely affected by certain environmental
factors. In the face of failure, for example, they suggest that low SEs will become
preoccupied with their deficiencies. In a related study, Cheek and Stahl (1986)
found that the poetry-writing creativity of nonshy women was unaffected by ex¬
pected external evaluation, but the creativity of shy women was significantly lower
under evaluation than under nonevaluation conditions.
These results on the linkage among performance, environmental conditions, and
self-esteem (or the related dimensions of self-consciousness and shyness) clearly
suggest a mediation mechanism: Environmental conditions such as expected evalua¬
tion can be perceived quite differently by persons who vary in self-esteem, and as a
result these environmental conditions can have quite disparate effects on the
creativity of persons who differ along this dimension.
Beyond whatever might be said about self-esteem as a mediator among environ¬
ment, motivation, and creativity, this line of inquiry suggests a new approach to
creativity research in general. For most of the past four decades, creativity re¬
searchers have focused almost exclusively on individual differences - the qualities
of talent, experience, and personality that distinguish highly creative persons from
their less creative peers. Our own research, and that of a few colleagues, has taken
the quite different approach of examining the influence of social-environmental
factors on intrinsic motivation and creativity. Both lines of inquiry have produced
interesting and useful information; both have contributed to theory and practice.
But, as is clear from our recent research and from work on individual difference
mediators of social effects, both lines of inquiry are incomplete.
There is no doubt that personal qualities of ability and personality have great
impact on creative behavior. There is no doubt that salient factors of extrinsic
constraint in the social environment can have a consistently negative impact on the
intrinsic motivation and creativity of most people most of the time. What we must
now develop are research paradigms acknowledging that neither class of factors, by
itself, can carry the day. Fluctuations in any individual’s level of creative output
must be examined in light of environmental influences on motivation, and environ-
Conditions of creativity 35

mental effects must be examined in light of an individual person’s perceptions of

these influences. Only then can the “conditions of creativity’’ be understood as
complex interactions between and among both internal and external conditions. We
believe that this is the crucial next step.

References
Amabile, T. M. (1979). Effects of external evaluation on artistic creativity. Journal of Personality and
Social Psychology, 37, 221-233.
Amabile, T. M. (1982a). Children’s artistic creativity: Detrimental effects of competition in a field
setting. Personality and Social Psychology Bulletin, 8, 573-578.
Amabile, T. M. (1982b). Social psychology of creativity: A consensual assessment technique. Journal of
Personality and Social Psychology, 43, 997-1013.
Amabile, T. M. (1983). The social psychology of creativity. New York: Springer-Verlag.
Amabile, T. M., DeJong, W., & Lepper, M. R. (1976). Effects of externally imposed deadlines on
subsequent intrinsic motivation. Journal of Personality and Social Psychology, 34, 92-98.
Amabile, T. M., & Gitomer, J. (1984). Children’s artistic creativity: Effects of choice in task materials.
Personality and Social Psychology Bulletin, 10, 209-215.
Amabile, T. M., Golfarb, P., & Brackfield, S. C. (1982). Effects of social facilitation and evaluation on
creativity. Unpublished manuscript, Brandeis University, Waltham, MA.
Amabile, T. M., & Gryskiewicz, S. S. (in press). Creative environment and personality: Innovation in
the R&D lab. In R. L. Kuhn (Ed.), Handbook for creative and innovative managers. New York:
McGraw-Hill.
Amabile, T. M., Hennessey, B. A., & Grossman, B. S. (1986). Social influences on creativity: The
effects of contracted-for reward. Journal of Personality and Social Psychology, 50, 14-23.
Barron, F. (1965). The psychology of creativity. In T. Newcomb (Ed.), New directions in psychology
(Vol. 2, pp. 3-134). New York: Holt, Rinehart & Winston.
Barron, F. (1968). Creativity and personal freedom. New York: Van Nostrand.
Bates, J. (1979). Extrinsic reward and intrinsic motivation: A review with implications for the class¬
room. Review of Educational Research, 49, 557-576.
Bern, D. (1972). Self-perception theory. In L. Berkowitz (Ed.), Advances in experimental social psy¬
chology (Vol. 6, pp. 1-62). New York: Academic Press.
Berglas, S., Amabile, T. M., & Handel, M. (1981). Effects of evaluation on children’s artistic
creativity. Unpublished manuscript, Brandeis University, Waltham, MA.
Boggiano, A. K., & Barrett, M. (1984). Performance and motivational deficits of helplessness: The role
of motivational orientations. Unpublished manuscript, University of Colorado, Boulder.
Brockner, J., & Hulton, A. J. (1978). How to reverse the vicious cycle of low self-esteem: The
importance of attentional focus. Journal of Experimental Social Psychology, 14, 564-578.
Bruner, J. (1962). The conditions of creativity. In H. Gruber, G. Terrell, & M. Wertheimer (Eds.),
Contemporary approaches to creative thinking (pp. 1-30). New York: Atherton.
Calder, B., & Staw, B. (1975). Self-perception of intrinsic and extrinsic motivation. Journal of Person¬
ality and Social Psychology, 31, 599-605.
Cattell, R. B., & Butcher, H. J. (1968). The prediction of achievement and creativity. New York:
Bobbs-Merrill.
Check, J. F. (1969). An analysis of differences in creative ability between white and negro students,
public and parochial, three different grade levels, and males and females. Final report. Wash¬
ington, DC: Office of Education.
Cheek, J. M., & Stahl, S. S. (1986). Shyness and verbal creativity. Journal of Research in Personality,
20, 51-61.
Condry, J. (1977). Enemies of exploration: Self-initiated versus other-initiated learning. Journal of
Personality and Social Psychology, 35, 459-477.
36 B. A. Hennessey and T. M. Amabile

Coopersmith, S. (1967). The antecedents of self-esteem. San Francisco: Freeman.

Crutchfield, R. (1961). The creative process. In Proceedings of the conference on ‘ ‘the creative person '
(pp. VI-l-VI-16). Lake Tahoe: University of California Alumni Center.
Crutchfield, R. (1962). Conformity and creative thinking. In H. Gruber, G. Terrell, & M. Wertheimer
(Eds.), Contemporary approaches to creative thinking (pp. 120-140). New York: Atherton.
Csikszentmihalyi, M. (1975). Beyond boredom and anxiety. San Francisco: Jossey-Bass.
deCharms, R. (1968). Personal causation. New York: Academic Press.
Deci, E. (1971). Effects of externally mediated rewards on intrinsic motivation. Journal of Personality
and Social Psychology, 18, 105-115.
Deci, E. (1972). Intrinsic motivation, extrinsic reinforcement, and inequity. Journal of Personality and
Social Psychology, 22, 113-120.
Deci, E. (1975). Intrinsic motivation. New York: Plenum.
Deci, E., Nezlek, J., & Sheinman, L. (1981). Characteristics of the rewarder and intrinsic motivation of
the rewardee. Journal of Personality and Social Psychology, 40, 1-10.
Deci, E., & Porac, J. (1978). Cognitive evaluation theory and the study of human motivation. In M.
Lepper & D. Greene (Eds.), The hidden costs of reward. Hillsdale, NJ: Lawrence Erlbaum.
Deci, E., & Ryan, R. (1980). The empirical exploration of intrinsic motivational processes. In L.
Berkowitz (Ed.), Advances in experimental social psychology. New York: Academic Press.
Deci, E., & Ryan, R. (1985). Intrinsic motivation and self-determination in human behavior. New York:
Plenum.
Feldman, D. (1980). Beyond universal in cognitive development. Norwood, NJ: Ablex.
Folger, R., Rosenfield, D., & Hays, R. P. (1978). Equity and intrinsic motivation: The role of choice.
Journal of Personality and Social Psychology, 36, 557-564.
Fromm, E. (1959). The creative attitude. In H. H. Anderson (Ed.), Creativity and its cultivation. New
York: Harper & Row.
Garbarino, J. (1975). The impact of anticipated reward upon cross-age tutoring. Journal of Personality
and Social Psychology, 32, 421-428.
Garwood, D. (1964). Personality factors related to creativity in young scientists. Journal of Abnormal
and Social Psychology, 68, 413-419.
Gergen, K. J. (1982). Toward transformation in social knowledge. New York: Springer-Verlag.
Getzels, J. W., & Csikszentmihalyi, M. (1976). The creative vision: A longitudinal study of problem-
finding in art. New York: Wiley-Interscience.
Greene, D., & Lepper, M. (1974). Effects of extrinsic rewards on children’s subsequent interest. Child
Development, 45, 1141-1145.
Guilford, J. P. (1950). Creativity. American Psychologist, 5, 444-454.
Harackiewicz, J. (1979). The effects of reward contingency and performance feedback on intrinsic
motivation. Journal of Personality and Social Psychology, 37, 1352-1363.
Harlow, H. (1950). Learning and satiation of response in intrinsically motivated complex puzzle perfor¬
mance by monkeys. Journal of Comparative Physiological Psychology, 43, 289-294.
Harter, S. (1978). Effectance motivation reconsidered: Toward a developmental model. Human Devel¬
opment, 21, 34-64.
Harter, S. (1981). A new self-report scale of intrinsic versus extrinsic orientation in the classroom.
Developmental Psychology, 17, 300-312.
Harter, S. (1982). The Perceived Competence Scale for Children. Child Development, 53, 87-97.
Hughes, T., & McCullough, F. (Eds.). (1982). The journals of Sylvia Plath. New York: Dial.
Hunt, J. (1965). Intrinsic motivation and its role in psychological development. In D. Levine (Ed.),
Nebraska Symposium on Motivation (Vol. 13). Lincoln: University of Nebraska Press.
Ickes, W. J., Wicklund, R. A., & Ferris, C. B. (1973). Objective self-awareness and self-esteem.
Journal of Experimental Social Psychology, 9, 202-219.
John-Steiner, V. (1985). Notebooks of the mind. Albuquerque: University of New Mexico Press.
Kamiol, R., & Ross, M. (1977). The effects of performance irrelevant rewards on children’s intrinsic
motivation. Child Development, 48, 482-487.
Conditions of creativity 37

Kemoodle-Loveland, K., & Olley, J. (1979). The effect of external reward on interest and quality of task
performance in children of high and low intrinsic motivation. Child Development, 50, 1207-4210.
Koestner, R., Ryan, R., Bemieri, F., & Holt, K. (1984). Setting limits on children’s behavior: The
differential effects of controlling vs. informational styles on intrinsic motivation and creativity.
Journal of Personality, 52, 233-248.
Kruglanski, A. (1975). The endogenous-exogenous partition in attribution theory. Psychological Re¬
view, 82, 387-406.
Kruglanski, A. (1978). Endogenous attribution and intrinsic motivation. In M. Lepper & D. Greene
(Eds.), The hidden costs of reward. Hillsdale, NJ: Lawrence Erlbaum.
Kruglanski, A., Friedman, I., & Zeevi, G. (1971). The effects of extrinsic incentive on some qualitative
aspects of task performance. Journal of Personality, 39, 606-617.
Lepper, M., & Greene, D. (1975). Turning play into work: Effects of adult surveillance and extrinsic
rewards on children’s intrinsic motivation. Journal of Personality and Social Psychology, 31, 479-
486.
Lepper, M., & Greene, D. (1978). The hidden costs of reward. Hillsdale, NJ: Lawrence Erlbaum.
Lepper, M., Greene, D., & Nisbett, R. (1973). Undermining children’s intrinsic interest with extrinsic
rewards: A test of the “overiustification” hypothesis. Journal of Personality and Social Psychol¬
ogy, 28, 129-137.
Lepper, M., Sagotsky, G., Dafoe, J., & Greene, D. (1982). Consequences of superfluous social
constraints: Effects on young children’s social inferences and subsequent intrinsic interest. Journal
of Personality and Social Psychology, 42, 51-65.
Lonky, E., & Reihman, J. (1980, September). Cognitive evaluation theory, locus of control, and
positive verbal feedback. Paper presented at a meeting of the American Psychological Association,
Montreal.
McGraw, K. (1978). The detrimental effects of reward on performance: A literature review and predic¬
tion model. In M. Lepper & D. Greene (Eds.), The hidden costs of reward. Hillsdale, NJ: Lawrence
Erlbaum.
McGraw, K., & McCullers, J. (1979). Evidence of a detrimental effect of extrinsic incentives on
breaking a mental set. Journal of Experimental Social Psychology, 15, 285-294.
Nicholls, J. G. (1972). Creativity in the person who will never produce anything original and useful: The
concept of creativity as a normally distributed trait. American Psychologist, 27, 717-727.
Pinder, C. C. (1976). Additivity versus non-additivity of intrinsic and extrinsic incentives: Implications
for theory and practice. Journal of Applied Psychology, 61, 693-700.
Pittman, T. S., Davey, M. E., Alafat, K. A., Wetherill, K. V., & Kramer, N. A. (1980). Informational
versus controlling verbal rewards. Personality and Social Psychology Bulletin, 6, 228-233.
Pittman, T. S., Emery, J., & Boggiano, A. K. (1982). Intrinsic and extrinsic motivational orientations:
Reward-induced changes in preference for complexity. Journal of Personality and Social Psychol¬
ogy, 42, 789-797.
Pritchard, R. D., Campbell, K. M., & Campbell, D. J. (1977). Effects of extrinsic financial rewards on
intrinsic motivation. Journal of Applied Psychology, 62, 9-15.
Reiss, S., & Sushinsky, L. (1975). Overjustification, competing responses, and acquisition of intrinsic
interest. Journal of Personality and Social Psychology, 31, 1116-1125.
Renzulli, J. S. (1973). New Directions in Creativity: Mark 1. New York: Harper & Row.
Rogers, C. (1954). Towards a theory of creativity. ETC: A Review of General Semantics, 11, 249-260.
Rosner, S., & Abt, L. (Eds.). (1970). The creative experience. New York: Grossman.
Ross, M. (1975). Salience of reward and intrinsic motivation. Journal of Personality and Social Psychol¬
ogy, 32, 245-254.
Ryan, R. M.,& Grolnick, W. S. (1984). Origins and pawns in the classroom: Self-report and projective
assessments of individual differences in children’s perceptions. Unpublished manuscript, University
of Rochester, Rochester, NY.
Sanders, S. J., Tedford, W. H., & Hardy, B. W. (1977). Effects of musical stimuli on creativity.
Psychological Record, 27, 463-471.
38 B. A. Hennessey and T. M. Amabile

Shapira, Z. (1976). Expectancy determinants of intrinsically motivated behavior. Journal of Personality

and Social Psychology, 34, 1235-1244.
Simon, H. (1967). Understanding creativity. In J. C. Gowan, G. D. Demos, & E. P. Torrance (Eds.),
Creativity: Its educational implications. New York: Wiley.
Simonton, D. K. (1980). Thematic frame, originality, and musical Zeitgeist: A biographical and trans-
historical content analysis. Journal of Personality and Social Psychology, 38, 972-983.
Smith, W. E. (1974). The effects of social and monetary rewards on intrinsic motivation. Unpublished
doctoral dissertation, Cornell University, Ithaca, NY.
Sobel, R. S., & Rothenberg, A. (1980). Artistic creation as stimulated by superimposed versus separate
visual images. Journal of Personality and Social Psychology, 39, 953-961.
Stasinos, D. (1984). Enhancing the creative potential and self-esteem of mentally handicapped Greek
children. Journal of Creative Behavior, 18, 117-132.
Stein, M. (1974-1975). Stimulating creativity (Vols. 1-2). New York: Academic Press.
Steinbeck, J. (1969). Journal of a novel. New York: Viking.
Sternberg, R. J. (1985). Cognitive development in the gifted and talented. In F. D. Horowitz & M.
O’Brien (Eds.), The gifted and talented: A developmental perspective. Washington, DC: American
Psychological Association.
Taylor, D. (1960). Toward an information processing theory of motivation. In D. Jones (Ed.), Nebraska
Symposium on Motivation: 1960. Lincoln: University of Nebraska Press.
Torrance, E. (1966). The Torrance Tests of Creative Thinking: Technical-norms manual. Princeton, NJ:
Personnel Press.
Torrance, E. P. (1968). Torrance Tests of Creative Thinking: Directions manual and scoring guide.
Princeton, NJ: Personnel Press.
Turner, R. G., Scheier, M. F., Carver, C. S., & Ickes, W. J. (1978). Correlates of self-consciousness.
Journal of Personality Assessment, 42, 285-289.
Wallace, D. (1985). Giftedness and the construction of a creative life. In F. D. Horowitz & M. O’Brien
(Eds.), The gifted and talented: Developmental perspectives. Washington, DC: American Psycho¬
logical Association.
White, R. (1959). Motivation reconsidered: The concept of competence. Psychological Review, 66,
297-323.
Part II

The role of the individual in creativity

■
Psychometric approaches
2 The nature of creativity as manifest in its
testing

E. Paul Torrance

Creativity defies definition

Creativity defies precise definition. This conclusion does not bother me at all. In
fact, I am quite happy with it. Creativity is almost infinite. It involves every sense -
sight, smell, hearing, feeling, taste, and even perhaps the extrasensory. Much of it
is unseen, nonverbal, and unconscious. Therefore, even if we had a precise concep¬
tion of creativity, I am certain we would have difficulty putting it into words.
However, if we are to study it scientifically, we must have some approximate
definition. There have been many attempts to define creativity. They all seem to
have something in common, and yet each is slightly different. Some are more
inclusive than others. Some are more precise, but no more approximate - perhaps
less. Some of them are boring and uninteresting, whereas others excite the imagina¬
tion and the senses. I would like to review a few of my attempts.

Attempts at definition

When I began my formal study of creativity at the University of Minnesota in 1958,

I realized that I must have a definition. I carefully reviewed the many existing
definitions and my own observations of creative behavior as I knew it. I found
several approaches, and I saw value in each.

Newness as a criterion

The production of something new is included in almost all of the definitions I found,
either explicitly or implicitly. Thurstone (1952) argued that it does not make any
difference whether or not society regards an idea as novel. He maintained that an act
is creative if the thinker reaches the solution in a sudden closure that necessarily
implies some novelty to the thinker. To Thurstone, the idea might be artistic,
mechanical, or theoretical. It might be administrative if it solves an organizational
problem. It might be a new football play, a clever chess move, or a new slogan.
Stewart (1950) shared Thurstone’s view on this issue, maintaining that creative
thinking may occur even though the idea produced may have been produced by
someone else at an earlier time. By this definition, creative thinking may take place

43
44 E. P. Torrance

in the mind of the humblest woman or in the mind of the most distinguished
statesman, artist, or scientist.
Stein (1953), contrary to Thurstone and Stewart, insisted that creativity must be
defined in terms of the culture in which it appears. To him, “novelty” or “new¬
ness” meant that the creative product did not exist previously in the same form.
Stein also believed that to be creative the novel work had to be accepted as tenable
and useful or satisfying by a group in time (Stein, 1953, p. 322). He hypothesized
that studies of creative persons may reveal a sensitivity to the gaps in knowledge
that exist in their own culture and that their creativity may be manifest in calling
attention to these gaps.

Creativity versus conformity

A number of investigators (Crutchfield, 1962; Wilson, 1956) have defined creativ¬

ity by contrasting it with conformity. In general, creativity has been seen as contrib¬
uting original ideas, different points of view, and new ways of looking at problems.
Conformity has been seen as doing what is expected, not disturbing or causing
trouble for others. Perhaps another way of viewing this phenomenon is best ex¬
pressed in Lefrancois’s definition: “. . . just as very low intelligence is stupidity,
so very low creativity is ordinariness” (1982, p. 264).
However, Starkweather (1976) saw the creative person as being neither conform¬
ing nor nonconforming, but as being free to conform or not conform, depending on
what is true, pleasing, good, or beautiful. My observations and research confirm
Starkweather’s conclusions.

True, generalizable, and surprising

Selye (1962), in his definition, requires that basic discoveries or creative contribu¬
tions possess to a high degree and simultaneously three qualities: “They are true not
merely as facts but also in the way they are interpreted, they are generalizable, and
they are surprising in the light of what was known at the time of the discovery”
(1962, p. 402).
H. H. Anderson (1959) also emphasized the search for the truth in his definition
of creativity. He was especially insistent that the creative environment provide
freedom for one to respond truthfully with one’s whole person as one sees and
understands the truth.
Bartlett (1958) employed the term “adventurous thinking,” which he defined as
“getting away from the main track, breaking out of the mold, being open to
experience, and permitting one thing to lead to another” (1958, p. 103).

Definitions involving process

Spearman (1930) saw creative thinking basically as a process of seeing or creating

relationships, with both conscious and subconscious processes operating. Accord-
Creativity as manifest in testing 45

ing to one of his principles, when two or more percepts or ideas are given, a person
may perceive them to be in various relations (near, far, the cause of, the result of, a
part of, etc.). Another principle held that when any item and a relation to it are
cognized, then the mind can generate in itself another item so related.
Ribot (1906) and others after him have emphasized the capacity of thinking by
analogy as the essential, fundamental element of creative thinking. He maintained
that the process of analogizing gives rise to the most unforeseen and novel combina¬
tions, but he warned that it produces in equal measure absurd combinations and very
original inventions. Recognizing the nonrational aspects of creative thinking, sever¬
al investigators have called attention to the exercise of discrimination or choice as a
part of the creative process. Barchillon (1961), for example, says that the thinking
processes involved in creation are of two kinds: cogito, to shake and throw things
together; and intelligo, to choose and discriminate from many different alternative
possibilities and then synthesize and bind together elements in new and original
ways. What he has in mind by cogito is apparently similar to what Kubie (1958)
conceptualizes as taking place in the preconscious system. The preconscious is able
to scan experiences and memories, to condense, to join opposites, and to find
relationships at speeds impossible to achieve in the conscious system. The resulting
intuitions, however, are not very precise and are subject to the primary process type
of thinking.
Wallas (1926) identified four steps in the creative process: preparation, incuba¬
tion, illumination, and revision. Apparently the process flows somewhat as follows:
First, there is the sensing of a need or deficiency, random exploration, and a
clarification or “pinning down” of the problem. Then ensues a period of prepara¬
tion accompanied by reading, discussing, exploring, and formulating many possible
solutions, and then critically analyzing these solutions for advantages and disadvan¬
tages. Out of all this comes the birth of a new idea - a flash of insight, illumination.
Finally, there is experimentation to evaluate the most promising solution for
eventual selection and perfection of the idea. Such an idea might find embodiment
in inventions, designs, scientific theories, improved products or methods, novels,
musical compositions, paintings, or sculptures.
Among those who have elaborated and refined Wallas’s conceptualization are
Osborn (1948), Patrick (1955), Pames (1962), de Bono (1967), Pames, Noller, and
Biondi (1977). In fact, one can detect the “Wallas process” as the basis for almost
all of the systematic, disciplined methods of training in existence throughout the
world today.

Mental abilities approach

J. P. Guilford (1956, 1959, 1960, 1986) has conceptualized creativity in terms of

the mental abilities involved in creative achievement. In his well-known structure of
intellect, he sees creative thinking as clearly involving what he categorizes as
divergent production. He defines divergent production as the generation of informa¬
tion from given information, where the emphasis is on variety of output from the
46 E. P. Torrance

same source (innovation, originality, unusual synthesis or perspective). Included in

the divergent thinking category are the factors of fluency, flexibility, originality,
and elaboration. He concludes, however, that creative thinking cannot be equated
with divergent thinking. He believes that sensitivity to problems and redefinition
abilities are also important in creativity. The redefinition abilities involve transfor¬
mations of thought, reinterpretations, and freedom from functional fixedness in
deriving unique solutions. Sensitivity to problems seems to be essential in getting
the creative thinking process in motion.

Levels of creativity

A few investigators have contended that it is necessary to consider levels of creativ¬

ity. I. R. Taylor (1959) sought to reconcile some of the apparent differences in
opinion concerning creativity by suggesting that we think of creativity in terms of
levels. He suggested the following five levels:

1. Expressive creativity, as in the spontaneous drawings of children.

2. Productive creativity, as in artistic or scientific products where there are restrictions
and controlled free play.
3. Inventive creativity, where ingenuity is displayed with materials, methods, and
techniques.
4. Innovative creativity, where there is improvement through modification involving
conceptualizing skills.
5. Emergenative creativity, where there is an entirely new principle or assumption
around which new schools, movements, and the like can flourish.

Taylor pointed out that many people have the fifth level in mind when they talk
about creativity. Because this fifth level is so rare, the lower levels usually have
been involved in most investigations regarding creative behavior. Taylor also ob¬
jected to the frequent confusion of creativity and prevalent interpretations of tradi¬
tional logic, scientific method, and intelligence. He maintained that fantasy associa¬
tions and relaxation for unconscious play are so essential for creative thought that
creativity cannot be subjected to the same interpretations as logic and scientific
method.
Inventive creativity has perhaps been subjected to more systematic definition and
specification of criteria than any other level. U.S. Patent Office criteria have been
spelled out, and manuals have been developed for evaluators (Cochran, 1955).
Newness itself is not enough. The thing must not have been known before and must
be useful. In general, one is considered to have invented something if one has
originated something of merit that others skilled in the art or science to which it
relates would not have thought of or would not have achieved if they had thought of
it. The invention should overcome some prior failure or some special difficulty,
offer something remarkable or surprising, overcome prior skepticism about possible
success, or meet an unfulfilled need.
Creativity as manifest in testing 47

Choice of a definition

At the initiation of my formal research on creativity I had to choose a research

definition. As my research got under way I added some “artistic” definitions that
illuminated my research definition and enabled me to communicate some of the
things I discovered that I could not communicate with words. Prior to my formal
research I had been engaged in research concerning U.S. Air Force survival train¬
ing. The Survival School was training men to respond creatively in emergencies and
extreme conditions. I now refer to the definition I used then as my “survival
definition.” Before proceeding to what I have learned from testing creativity, I
would like to review these definitions.

My research definition

I chose a process definition of creativity for research purposes. I thought that if I

chose process as a focus, I could then ask what kind of person one must be to
engage in the process successfully, what kinds of environments will facilitate it, and
what kind of products will result from successful operation of the process (Tor¬
rance, 1965).
I tried to describe creative thinking as the process of sensing difficulties, prob¬
lems, gaps in information, missing elements, something askew; making guesses and
formulating hypotheses about these deficiencies; evaluating and testing these guess¬
es and hypotheses; possibly revising and retesting them; and finally communicating
the results.
I like this definition because it describes such a natural process. Strong human
needs appear to be at the basis of each of its stages. If we sense an incompleteness,
something missing or out of place, tension is aroused. We are uncomfortable and
want to do something to relieve the tension. As a result, we begin investigating,
asking questions, manipulating things, making guesses or hypotheses, and the like.
Until these hypotheses have been tested, modified, and retested, we are still uncom¬
fortable. Then, even when this is done, the tension is unrelieved until we tell
someone what we have discovered or produced. Throughout the process there is an
element of responding constructively to existing or new situations, rather than
merely adapting to them. Such a definition places creativity in the realm of every¬
day living and does not reserve it for ethereal and rarely achieved heights of
creation.
To illustrate the process, I sometimes give an audience as it assembles one of the
tests we have developed for assessing the creative thinking abilities, such as the
Incomplete Figures Test. After a while I check with the audience to find out what
has happened. Most of them will admit that the incompleteness or some other
quality of the figures made them uncomfortable. Usually some of them went ahead
and completed the figures in some way, either by actually drawing lines or imagina¬
tively by thinking of completing the figures. I then ask them to go ahead and
48 E. P. Torrance

complete the figures in some way that will be satisfying. There is then an obvious
atmosphere of relief, increased liveliness, even smiles and laughter. There is also
spontaneous interest in communicating the results and seeing what others have
created.
If the thinking of the audience is bound to some specific curricular content, such
as reading, art, science, or language arts, I try to show how the process operates
with that specific curricular content. For example, take the problem of reading
creatively. When one reads creatively, one must first of all become sensitive to the
problems and possible gaps in whatever one reads. One makes oneself aware of the
gaps in information, unsolved problems, missing elements, and things that are
incomplete or out of focus. To resolve this tension, the creative reader sees new
relationships, creates new combinations, synthesizes relatively unrelated elements
into coherent wholes, and redefines or transforms certain pieces of information to
discover new uses and build onto what is known. In this search, the creative reader
produces a variety of possibilities, uses many approaches, looks at the information
from many perspectives, breaks away from commonplace solutions into bold new
ways, and develops ideas by filling in the details and making these ideas attractive
or exciting to others.
If I want to communicate something about the qualities of the creative product, I
ask my audience to discuss the qualities of the products they have just produced. In
addition to divergent thinking qualities (fluency or number of ideas, flexibility or
shifts in approaches, originality or unusualness, and elaboration or amount of detail
or completeness manifested), we talk about such qualities as humor, fantasy, color¬
fulness and richness of imagery, unusual visualization, internal visualization,
boundary pushing, movement, articulateness in telling a story, and the like.
To develop an understanding of the importance of the “press” or environment, I
have on several occasions conducted a tape-recorded exercise created by Cun-
nington and myself (1962, 1965) called “Sounds and Images.” It has built into it
several “presses,” applications of research findings that have been shown to facili¬
tate originality and other qualities of creative thinking. The tape recording consists
of four different sound effects, ranging from an easily recognizable familiar sound
having few “missing elements” to a strange effect made up of six relatively
unrelated sounds. This set of four sound effects is presented with a slight pause
between each, with instructions to the listener to write down a word picture of the
image generated by each sound effect. The sound effects are then repeated, with
instructions to let the imagination roam more widely and depart from the most
obvious images. After this, the sound effects are repeated a third time, with instruc¬
tions to let the imagination swing free. An attempt is made to free the audience from
threats of evaluation and to encourage them to have fun. An attempt is also made to
break the “set” between each of the repetitions of the sound effects. Finally, the
audience is asked to select its favorite images and translate them into a picture. I
then ask members of the audience to describe what they have experienced, trying to
identify what features of the environment (the taped exercise, the situation, the
other members of the audience, etc.) facilitated or hindered their efforts to produce
Creativity as manifest in testing 49

CREATIVITY IS...

original images. They quickly identify such things as the built-in warm-up process,
going from easy to difficult, going from simple to complex, the legitimacy of
thinking divergently, freedom from evaluation, and the like (Torrance, 1965).
If I want to develop understandings about the importance of the person, I identify
built-in environmental features and then ask the audience to identify those forces in
themselves that facilitated or inhibited them in making use of these built-in helps.
This usually results in a rather long list of personality factors (most of which have
been found through research to be related to creative behavior), and hopefully in an
increased awareness by the audience of their own creative potential and an increased
understanding of the forces in individual personalities that influence their creative
functioning and that of their associates.

My “artistic” definition

Perhaps even more useful than my research definition has been my “artistic”
definition. It has been especially useful in generating hypotheses, suggesting ideas,
theorizing, organizing my thinking, and communicating the nature of creativity. It
was given to me in 1964 by Karl Anderson, a student of mine at the University of
California at Berkeley. It consisted of simple line drawings and simple sentences. I
long ago lost the original, but several artists have attempted to interpret the verbal
part of the definition. Most of their productions have been more elaborate than those
of Karl Anderson, and most of them have been more colorful, but none of them
have been more meaningful.
To make them really meaningful, one has to look on them as analogies. I suggest
that instead of saying of each, “Creativity is . . . ,” one say, “Creativity is
like. ...” Beginning with this as an analogy, one can play with it, as in the
“synectics” (Gordon, 1961) approach to creative problem solving, and elaborate it.
The drawings made by Nancy Martin in 1985 are more like the originals than any of
the others, and so I present them (Figures 2.1-2.14). I like the simple line draw-
CREATIVITY IS

Pigging Peeper
Figure 2.2.

CREATIVITY IS...

LOOKING TWICE

Figure 2.3.

CREATIVITY IS...

LISTENING FOL
5MELLS

Figure 2.4.
 CREATIVITY IS..

TAbKMr LISTENING*
TO A CAT

glossing oor Mistakes

Figure 2.5.

CREATIVITY IS...

Getting w

<$
Figure 2.6.

CREATIVITY IS

Getting out

Figure 2.7.
52 E. P. Torrance

CREATIVITY IS...

HAVING A tALL

CREATIVITY IS...

Cutting Holts to set through

Figure 2.9.

ings. Just enough detail is given to inspire an infinity (almost) of meaning and ideas.
Some of the other illustrations give more detail and are better for transmitting
information. For example, Matt Daly of Palmyra, New York, did an excellent set of
illustrations (Goetzmann, 1979) that communicate more information than those of
Karl Anderson and Nancy Martin. Consider Figure 2.15, which parallels Figure
2.11 (plugging into the sun). Figure 2.11 is very simple. It is not too difficult to
“discover” the meaning, but it is difficult enough to elicit “Aha!” from most
people, at least a small one. One does not find that in Figure 2.15. The analogy is
quite direct, saying “plug in for more energy,” and suggests seven sources. Of
course, a teacher or leader might say that these are only seven examples of sources
Creativity as manifest in testing 53

?Hl&&INQ tHTO THE sun

Figure 2.11.

of energy or information. The viewer has already been limited. There is not the
infinity of sources suggested by Figure 2.11, which uses the sun as the analogy.
What about religion, the home, music, art, exercise, ocean waves, the ocean’s roar,
and so forth, as sources of energy?
Let us also examine Figure 2.16, which parallels Figure 2.2 (digging deeper). It
is much more difficult to attribute meaning to Figure 2.16 than to Figure 2.2, and
when we do, we cannot be certain it is what the artist intended. Figure 2.16 was
drawn by Earl Ginter (1980) as a part of his creative self-exploration, trying to
54 E. P. Torrance

CREATIVITY IS...

Ym

recapture some of the skill and some of the love he had once had for art. Its potential
for generating ideas and concepts is great, but the risk is also great.
I like Figure 2.17, Matt Daly’s illustration for this analogy (Goetzmann, 1979).
He pictures several shallow holes that have been abandoned, and the digger now has
hit “pay dirt” (diamonds) by digging deeper. This analogy, however, has the
opposite result from that used by Edward de Bono (1967) in trying to communicate
his concept of lateral thinking (what I define as “creative thinking”). He explains
that vertical (logical) thinking digs the same hole deeper, whereas lateral thinking is
concerned with digging a hole in a different place. Young Matt Daly’s analogy
 CREATIVITY IS...

SHAKING HANDS WITH Tomorrow

Figure 2.15. Plug in for more energy.

56 E. P. Torrance

Figure 2.16. Creativity is digging deeper.

suggests that we may have to dig several or many holes before we settle on one
place and dig deeper.
I have for a long time particularly liked the following analogical definition of
creativity, in spite of the fact that it is entirely verbal:
Creativity: an arbitrary harmony, an expected astonishment, a habitual revelation, a familiar
surprise, a generous selfishness, an unexpected certainty, a formable stubbornness, a vital
triviality, a disciplined freedom, an intoxicating steadiness, a repeated initiation, a difficult
Creativity as manifest in testing 57

Figure 2.17. Dig deeper.

delight, a predictable gamble, an ephemeral solidity, a unifying difference, a demanding

satisfier, a miraculous expectation, an accustomed amazement. (Prince, 1970, p. xiii)

This definition is difficult or puzzling enough to bring forth some “Ahas” as

insights begin to be discovered. In teaching, it is useful to ask students to think of
examples of each analogy from their own lives or observations. Notice how difficult
it is to transform them into words once insights have occurred.

My “survival” definition

My briefest and in some ways most satisfactory definition of creativity is what I

refer to as my survival definition: When a person has no learned or practiced
solution to a problem, some degree of creativity is required. I spent 7 years directing
a program of research in support of U.S. Air Force survival training (1951-1957).
58 E. P. Torrance

The air force was training its aircrewmen to survive emergencies and extreme
conditions (cold, heat, lack of food and/or water, lack of shelter, lost at sea or in the
jungle, down in enemy territory, etc.)- They were given information about how they
might deal with all these environmental conditions. They were given information
about how others had escaped from POW camps and successfully evaded the
enemy. In survival training, crews were also practiced in simulated situations.
However, in the actual emergency and extreme conditions, the aircrewman was
facing a new situation for which he had no learned and practiced solution.
The truly creative is always that which cannot be taught. Yet creativity cannot
come from the untaught. Creative solutions to aircrew survival situations required
imaginatively gifted recombination of old elements (information about how the
American Indians had lived off the land, how the early explorers survived in the
Arctic, how men had survived shipwreck, how airmen in World Wars I and II had
escaped and evaded, etc.) into new configurations - what is required now. The
elements of a creative solution can be taught, but the creativity itself must be self-
discovered and self-disciplined. Thus, an important part of survival training in¬
volved practicing means of self-discovery and self-discipline and use of the imag¬
ination (Torrance, 1957a).

The case for learning about the nature of creativity from testing

My chief way of learning about the nature of creativity has been through testing and
teaching creative behavior. I believe that these are valid ways of knowing and that
testing and teaching creative behavior have analogues in real-life creative achieve¬
ments. The strongest evidence of a relationship between test behavior and real-life
creative achievement comes from two longitudinal studies: one involving high
school students tested in 1959 and followed up 7 and 12 years later and the other
involving elementary school pupils tested in 1958 and for 5 subsequent years and
followed up 22 years later (Torrance, 1972a, 1972b, 1981).
The criteria of creative adult behavior consisted of the following indices (Tor¬
rance, 1981):
1. Quantity of publicly recognized and acknowledged creative achievements (patents
and inventions; novels; plays that were publicly produced; musical compositions
that were publicly performed; awards for artworks in a juried exhibition; founding a
business; founding a journal or professional organization; developing an innovative
technique in medicine, surgery, science, business, teaching, etc.).
2. Quality of creative achievements. Subjects were asked to identify what they consid¬
ered their three most creative achievements; these data, plus responses to the
checklist of achievements, were judged by three judges.
3. Quality of creative achievement implied by future career image. Three judges
assessed this primarily by responses to the following two questions in the follow¬
up: (1) What are your career ambitions? For example, what position, responsibility,
or reward do you wish to attain? What do you hope to accomplish? (2) If you could
do or be whatever you choose in the next 10 years, what would it be?
4. Quantity of high school creative achievements (used only in elementary school
study) (similar to item 1, but limited to achievements during the high school years).
5. Quantity of creative style of life achievements, not publicly recognized (such as
Creativity as manifest in testing 59

Table 2.1. Product-moment coefficients of correlation between creativity

predictors established in 1959 and the criterion measures established
in 1966 and 1971

Criterion variables

Quality Quality Quantity Quantity Aspiration Aspiration

1966 1971 1966 1971 1966 1971
Predictions N = 46 N = 52 N = 46 N = 52 N = 46 N = 52

Fluency (TTCT)a .39* .53* .44* .54* .34 .49*

Flexibility (TTCT) .48* .59* .44* .58* .46* .54*
Originality (TTCT) .43* .49* .40* .54* .42* .51*
Elaboration (TTCT) .32 .40* .37* .43* .25 .41*

“Torrance Tests of Creative Thinking.

^Coefficient of correlation is significant at the .01 level.

Table 2.2. Product-moment coefficients of correlation

between predictor and criterion measures of total
high school sample

Predictors Number Quantity Quality Aspiration

Fluency 254 .30“ .30“ .27“

Flexibility 254 .28“ .29“ .24“
Inventiveness 254 .36“ .41“ .35“
Total (1959) 254 .32“ .36“ .32“
Originality 254 .40“ .43“ .39“
Intelligence 254 .21“ .38“ .39“
Achievement 254 .27“ .47“ .43“

“r is significant at better than the .01 level.

organizing an action-oriented group, designing a house, designing a garden, initiat¬

ing a new educational venture, painting that was not exhibited, musical composi¬
tion not publicly performed, etc.) (used only in elementary school study).

In the high school group, follow-up data were obtained from the seniors in 1966
and 1971 (N = 46 in 1966 and 52 in 1971). In 1971, criterion data were obtained for
254 altogether. In 1980, criterion data were obtained from a total of 211.
Table 2.1 shows the results obtained from the seniors in the high school study
who supplied follow-up data in 1966 and 1971 (Torrance, 1972a). It will be ob¬
served that the creativity test predictors successfully predicted the criterion mea¬
sures (judged quality of three most creative achievements, number of publicly
acknowledged and recognized creative achievements, and judged quality of creative
achievements projected for the future) in 1966 and 1971. It is interesting to note that
the predictions are somewhat higher in 1971 than in 1966.
60 E. P. Torrance

Table 2.3. Predictive validity of the Torrance Tests of Creative Thinking

for males, females, and total sample for five criteria of creative achievement
in 22-year follow-up

Males Females Total

Criteria of creative achievement (N = 95) (N = 116) (N = 211)

Number of High School Creative Achievements .33 .44 .38

Number of Post-High-School Creative
Achievements .58 .42 .46
Number of Creative Style-of-Living Achievements .42 .48 .47
Quality of Highest Creative Achievements (ratings) .59 .57 .58
Quality of Future Career Images .62 .54 .57

Note: All coefficients of correlation are significant at the .001 level.

The predictive validity for the entire high school sample (N — 254) supplying
follow-up data is shown in Table 2.2 (Torrance, 1972a, 1972b). Again, it is seen
that all of the creativity test predictors are statistically significant at the .01 level.
The combined predictor correlates with the combined criterion at .51 (.59 for males
and .46 for females).
From these studies (Torrance, 1972a, 1972b) the following conclusions were
drawn:
1. Young people identified as creative on the basis of creativity tests during the high
school years tend to become productive, creative adults.
2. At least 12 years after high school graduation appears to be a more advantageous
time than 7 years as the time for a follow-up of creative adults.

For the 22-year follow-up of the elementary school students first pretested in
1958, the predictive validity is shown in Table 2.3. Again, all the creativity test
predictors correlate with all of the creativity achievement criteria at the .001 level or
higher. When all of the predictors are combined to predict the combined criterion, a
correlation of .63 is attained. For males, the overall validity coefficient was .62; for
females, it was .57. Although the creativity test predictors leave considerable unex¬
plained variance, it is unusual to find higher predictive validity for intelligence and
achievement tests or other predictor variables in similar studies. In the present
study, validity coefficients for measures of intelligence ranged from —.02 to .34
and averaged . 17. In the present study it was found that additional variance could be
explained by such things as having certain teachers known for encouraging creativ¬
ity, having a mentor, having a future career image during the elementary school
years, and having experiences with foreign study and living. Thus, I believe that the
predictive validity is as good as we have any right to expect for almost any kind of
predictor of adult achievements (Torrance, 1981).
Although there have been several predictive validity studies with the Torrance
Tests of Creative Thinking (TTCT), as well as with other creativity predictors,
Creativity as manifest in testing 61

Table 2.4. Product-moment coefficients of correlation between Sounds and

Images scores in 1961 and criteria of creative achievement in 1980 (N = 92)

Scoring variable Quality Quantity Future Style

Originality of images .31" .35" .39" .40"

Strangeness, fantasy .32" .31" .28" .33"
Unusual sensory images .25" .27" .26* .25*
Synthesis into coherent whole .35" .38" .29" .45"
Colorful image, rich detail, appealing .40" .44" .34" .31"
Movement and action images .25* .25* .17 .14
Number of correct sounds as images -.14 — .21* -.09 -.14"
Number of common images -.30" -.27" -.40" -.32"

"Significant at .01 level.

^Significant at .05 level.

except for two by Howieson (1981, 1984), all of them have been for relatively short
periods (usually about 5 years). Howieson (1981) reported a 10-year follow-up
study covering the period from 1965 to 1975 in Australia. Her subjects were 400
seventh graders, and the criterion data consisted of responses to the Wallach and
Wing (1969) checklist of creative achievements outside of the school curriculum.
The total score on the TTCT correlated .30 with the total criterion score. As in the
Torrance studies, the predictions were more accurate for the males than for the
females. Howieson (1984) completed a 23-year follow-up study using predictor
data collected in 1960 by Torrance in Western Australia. Although the verbal
measures derived from the TTCT failed to predict adult creative achievements at a
satisfactory level for the 306 subjects who returned their questionnaires, the figural
measures fared rather well. For her measure of quality of publicly recognized
creative achievements, Howieson obtained a multiple correlation coefficient of .51;
for quantity of personal (not publicly recognized) creative achievements, a multiple
coefficient of correlation of .33; and for quality of personal creative achievements,
one of.44.
Sounds and Images (Khatena & Torrance, 1973), another of the TTCT, was also
used as a predictor in the elementary follow-up study. Sounds and Images was
administered in 1961 to 92 of the subjects who supplied follow-up data. The results
of this study are shown in Table 2.4 (Torrance, 1984). From the data presented in
Table 2.4 it will be observed that almost all of the hypothesized relationships were
statistically significant. Originality in images, strangeness, unusual sensory images,
coherent syntheses, colorful images, and movement and action images were posi¬
tively related to the criteria of young adult creative achievement. Only two of the
coefficients of correlation, for movement and action images (Future Career Image
and Creative Style-of-Living Achievements), failed to meet the criterion of statis¬
tical significance (p < .05). The number of common images (zero scores) was
consistently correlated negatively with the criteria. A similar trend was found for a
62 E. P. Torrance

number of specific sounds correctly identified and reported as images, but only the
correlations for Quantity of Creative Achievements and Number of Creative Style-
of-Living Achievements reached statistical significance.
Although several hundred studies have dealt with the validity of creativity tests,
these longitudinal studies with real-life criteria seem to offer the strongest link to
test behavior of creative achievement. We seem to be justified in assuming analo¬
gies between test behavior and school or training behavior and “real-life” creative
behavior.

Manifestations of the nature of creativity through creativity testing

I believe that I have learned some useful things about the nature of creativity by
studying variations in testing conditions, the skills involved in creative thinking, the
characteristics of the creative person, and group factors that facilitate or interfere
with creative function. I shall review briefly each of these areas.

Variations in test conditions

Creativity tests, especially the TTCT, have been subjected to more experimentation
regarding ways of administering them than any other educational or psychological
tests in the history of testing. This has been good. From it, we have learned a great
deal about what procedures enhance, facilitate, or hinder creative thinking in gener¬
al. One conclusion is clear: How a creativity test is administered, and under what
conditions, can influence performance. Also, rather consistently, such experiments
have shown that the test administration instructions given in the manual do give the
most reliable and valid results.
Even before publication of the TTCT, I had performed numerous experiments
concerning the wording of the instructions for these tests and had made systematic
observations regarding various conditions. For example, we found that addition of
the words “Try to think of something no one else will think of” decreased the
amount of copying from one another and increased the originality of responses. I
observed that children in an overheated, stuffy classroom were poorly motivated
and did not perform as well as children in a comfortable and well-ventilated room. I
also observed that children in late May did not perform as well as they had per¬
formed earlier, despite training and motivation of high quality.
After publication of the tests in 1966,1 was able to locate 16 experiments that had
been published between 1968 and 1972 involving variations in testing conditions.
They are summarized Table 2.5 (Torrance, 1972c). In 1982, I made a similar
survey of studies that had been published between 1972 and 1982. Twenty such
experiments were identified and are summarized in Table 2.6. A detailed analysis of
these results indicates that for 70% of the measures, the experimental testing condi¬
tion did result in statistically significant increases. From them, we learn much that is
useful about maintaining classroom conditions favorable to learning and creativity.
Some of them might be incorporated in standard testing conditions. Some are
Table 2.5. Summary of experiments involving variations in testing conditions
prior to 1972 survey

Grade Significant (.05)

Investigator level Nature of treatment differences

Aliotti (1969) 1 Movement and verbal TTCT differences not

warm-up day prior to significant
testing
Boersman & O’Bryan 4 Standard vs. relaxed Relaxed: TTCT
(1968)
Elkind et at. (1970) 5-12 years Interruption of interesting Uninteresting tasks signif¬
vs. uninteresting task icantly higher on Wal-
lach-Kogan test
Feldhusen et al. (1971) 5, 8 Standard, incubation, Highest TTCT r’s with
take-home, game-like achievement on standard
and lowest on game-like
Hooper & Powell (1971) 1, 3 Absolute music vs. pro¬ Absolute music, TTCT
gram music
Khatena (1971) 10, 12 Variations in time limits Increased time for in¬
for response cubation, increased origi¬
nality, TTCT
Kogan & Morgan (1969) 5 Test-like and game-like Game-like, higher fluen¬
cy and (timed) unique re¬
sponses Wallach tests
Mohan (1970) 4 Cue-rich and cue-poor TTCT variables; helped
testing room high creatives more than
lows
Nash (1971) 1 Warm-up immediately TTCT Figural increased
prior to testing significantly
Norton (1971) 6 Music No significant differences
TTCT
Roweton & Spencer Intermediate Practice Significant effect only on
(1972) Figural A, TTCT
Torrance (1969) 6 (gifted) Take-home after timed Take-home more valid
administration for teacher curiosity nom¬
inations, TTCT
Towell (1972) 4 Untimed No significant increment,
TTCT
Van Mondfrans el al. 5, 8 Standard, incubation, Standard, highest verbal
(1971) take-home, game-like means; take-home, scores
that fit best concept of
creativity as unitary fac¬
tor orthogonal to intel¬
ligence, TTCT
Ward (1969) Nursery Cue-poor, cue-rich No significant environ¬
environment ment effect Wallach-
Kogan measures
Ward (1969) 7-8 years Successive time periods Increased uncommonness
with time

Source: Data from Torrance (1972c, pp. 131-132).

Table 2.6. Summary of experiments involving variations in testing conditions
reported between 1972 and 1982
Grade
Investigator level Nature of treatment
Aliotti (1973) 1 Individual administration,
verbal warm-up, physical
warm-up
Belcher (1975) 4, 5 (1) Saw fdm of model
giving original; (2) same
model giving unoriginal
responses; (3) read book
designed to stimulate
ideation
Carroll (1980) 2, 4, 6 Warm-up, peer evalua¬
tion, teacher evaluation
Dixon (1973) 5, 6 Evaluation vs. unevalua¬
tion atmosphere
Friedman, Raymond, Feld- (1) Barren room, (2) en¬
husen (1978) riched room, (3) look
around for cues, (4) en¬
riched room + look
around for cues each 5
min
Frost (1976) 5 Frustrating and non¬
frustrating conditions
Frost (1976) 11, 12 Three levels of frustration
Gakhar (1974) 7-9 Standard conditions, en¬
couraged to be more origi¬
nal, untimed
Hattie (1980) 6 (1) Untimed, game-like,
(2) timed, test-like, (3)
test-like, 2 consecutive
days
Hudgins (1974) HS High and low structure
Johnson (1974) 3, 5 Immediate reward, de¬
layed reward, no reward
Kaltsounis (1973) 5 Sound, speech, music,
silence
Kandil (1980) 1-5 (EH) Positive verbal reinforce¬
ment vs. no verbal rein¬
forcement, individual
administration
Significant (.05)
differences
TTCT Verbal, no
difference
Both film groups better on
TTCT Verbal fluency &
Originality; group 1
excelled group 2
TTCT no difference in 2
and 4; warm-up group sig¬
nificantly better in 6
TTCT Verbal, no
difference
TTCT (2 Figural, 1 Ver¬
bal), significantly higher
for enriched room + in¬
structor to keep looking
TTCT, no significant
difference
TTCT Figural, no
difference
TTCT significantly high¬
er, encouraged & untimed
TTCT, generally higher
on both test-like
conditions
TTCT Lines and Circles,
significantly higher syn¬
thesis for low structure
TTCT Figural, reward
conditions consistently
significantly higher
TTCT Figural, music
group significantly higher
except for elaboration
TTCT Verbal, consistent
significant differences on
all measures in favor of
positive verbal
reinforcement
Creativity as manifest in testing 65

Table 2.6 (Continued)

Grade Significant (.05)

Investigator level Nature of treatment differences

Kerns (1979) 6, 10 Unisensory vs. bisensory TTCT Figural, signifi¬

stimulation
Landreneau & Halpin 6 Modeling of response to
(1978) unusual uses, varying lev¬
els of flexibility and
originality
Lero (1974) 5 Timed vs. untimed
McCormack (1975) 3 Nonverbal vs. verbal
directions
Peters & Torrance (1972) 2-6 years Same-sex dyads vs. indi¬
vidual administration &
then same-sex dyads
White & Torrance (1978) 2 Warm-up with song How
Many Ways vs. no warm¬
up
Ziv (1976) 10 Humor (played record of
comedian) vs. standard
conditions
cantly higher on all mea¬
sures for unisensory
TTCT (Just Suppose &
Unusual Uses), positive
effect only on Fluency
TTCT Lines and Circles,
untimed group signifi¬
cantly higher on Fluency
and Originality
TTCT Figural B, signifi¬
cantly higher Fluency &
Elaboration for nonverbal
Construction task with
blocks, dyads enhanced
performance of 6-year-
olds; no effects on 2-5
TTCT Circles, signifi¬
cantly higher for warm-up
TTCT Unusual Uses &
Just Suppose, significant
difference in favor of
humor

impractical insofar as tests go; yet they certainly can be used as a part of teaching.
For example, it might not be practical and fair as a test administration practice to
give creativity tests as take-home tests. Yet we try to keep the incubation processes
going in every lesson.
Although the results of these two surveys seem to show some inconsistencies,
there also are some clear and consistent trends. Generally, psychological warm-up
given before testing results in small but consistent and statistically significant gains
over standard conditions. Warm-up that is too long or that is given too far in
advance of the testing does not help. Doing something to set the incubation pro¬
cesses to work ahead of time can be helpful.
Take-home testing results in greater variability. In most instances, scores are no
higher than under timed conditions. However, if one is successful in motivating
students and in setting up conditions for incubation, take-home tests are bound to
produce superior results.
Both relaxed and stressful conditions can produce increased performance. Some
arousal is necessary for creativity, but too much stress is detrimental. In the same
66 E. P. Torrance

way, conditions may become too playful, so that no work is accomplished, and
making creative productions is “work.” Students need to be made comfortable,
physically and psychologically, so that they can focus on the creative task.
It is important, however, to give some reason for the testing. It should make
sense. The subjects should be told why they are being tested. Some creativity tests
do not do this; subjects are deliberately not told what they are being tested for. If we
wanted to measure the jumping ability of the children in our school or the young
people in a college, we would not measure how high they just happened to be
jumping. We would do something to get them jumping. We would also tell them if
we wanted to see how high they could jump, how far they could jump, and so forth.
It is necessary to do the same thing to get a valid measure of creative thinking
ability. For this reason, subjects are motivated on the TTCT to fluency, flexibility
originality, elaboration, and so forth.
The experiments on providing cue-rich and cue-poor environments for testing
also proved to be productive. Mohan (1971) found that children scored higher in a
room rich in cues (a variety of objects in view in the room) than did those in a barren
room. However, he found that this difference came almost entirely from the high
creatives. The cue-rich environment did not help the low creatives; apparently they
were not accustomed to scanning the environment for cues to ideas for transforma¬
tion, combinations, and the like. However, Friedman, Raymond, and Feldhusen
(1978) found that by suggesting every 5 min that the subjects look around, higher
scores could be achieved. Deliberate training of low creatives to scan the environ¬
ment and instruction in how to use environmental cues might be used to improve
their creative functioning.

Skills involved in creative thinking

Edward de Bono (1975) sees thinking as a skill like reading, writing, riding a
bicycle, swimming, playing tennis, cooking, or skiing. Improvement in thinking
skills requires practice with the right tools and in the right environment. The
problem then becomes one of what component skills underlie creative thinking and
how these can best be trained.
On the TTCT (Torrance, 1979, 1984) we have identified a variety of skills that
seem to be important in producing creative responses. They are mentioned fre¬
quently in studies of the creative giants, in personality studies of creative persons, in
training guides, and in other creativity literature. The responses to both the verbal
and figural tests, Sounds and Images, and other creativity tests give evidence of
these abilities, which can be scored (Torrance, 1981, 1984; Wechsler, 1982). We
have also shown that they are valid predictors of creative achievement.
For example, we have identified the following abilities in the figural forms of the
TTCT (Torrance, 1979; Torrance & Ball, 1984):
1. Fluency, number of responses.
2. Flexibility, number of ways circles or lines were used.
3. Originality, unusualness, or rarity of the response.
Creativity as manifest in testing 67

4. Elaboration, number of details that contribute to the “story” told by the response.
5. Abstractness of the title, level of abstraction of the response.
6. Resistance to closure on the incomplete figures or the ability to “keep open.”
7. Emotional expressiveness of the response.
8. Articulateness of storytelling, putting the response in context, giving it an en¬
vironment.
9. Movement or action shown in the response.
10. Expressiveness of the titles, ability to transform from the figural to the verbal and
give expression.
11. Synthesis or combination, joining together two or more figures and making it into a
coherent response.
12. Unusual visualization, seeing and putting the figure in a visual perspective different
from the usual.
13. Internal visualization, seeing objects from the inside.
14. Extending or breaking the boundaries, getting outside the expected.
15. Humor, juxtaposition of two or more incongruities.
16. Richness of imagery, showing variety, vividness, liveliness, and intensity.
17. Colorfulness of imagery, exciting, appeal to the senses, flavorful, earthy, emo¬
tionally appealing.
18. Fantasy, unreal figures, magic, fanciful fairy tale characters, science fiction
characters.

In order to validate these skills as represented by test scores, I had the test
protocols of samples from my two longitudinal studies rescored by independent
scorers according to the manual (Torrance & Ball, 1984) and correlated with two
criteria: judged quality of creative achievements and number of publicly recognized
and acknowledged creative achievements. In the high school study (Torrance &
Ball, 1984), all of the correlations were statistically significant, ranging from .22
to .79 (N ~ 162) and averaging .53. An index made up of all the variables cor¬
related .79 with the combined criteria. In the elementary school study with the
criteria collected 22 years later, the total index correlated with the criterion .50 for
tests taken in the third grade, .34 for the fourth grade, .46 for the fifth grade,
and .42 for the sixth grade.

Characteristics of the creative person

I titled the volume containing a collection of my papers about the creative person
The Blazing Drive (Torrance, 1987). This title sums up the essence of what I have
learned about the nature of creativity and the creative person. The title was inspired
by Jonathan Livingston Seagull (Bach, 1970), in which Jonathan says of his prize
pupil, “He was strong and light and quick in the air, but far and away more
important, he had a blazing drive to learn to fly” (p. 75). Teresa Amabile (1986)
recently wrote that “extraordinary talent, personality, and cognitive ability do not
seem to be enough - it’s the ‘labor of love’ aspect that determines creativity”
(p. 12). This conclusion comes from dozens of findings from testing for creativity,
but especially from my longitudinal studies.
In my 22-year study of elementary school children first tested in 1958, I asked
them each what they were in love with, what they wanted to become when they
68 E. P. Torrance

were adults (Torrance, 1981). Some of the children consistently responded “I don’t
know.” Others were inconsistent, changing their future images every year. Most of
these continued to change and are now in work different from any that they had
mentioned as youngsters. Some gave responses in terms of what was expected or
what they thought was expected of them. For example, one second grade boy’s
response was “physician or surgeon,” the occupation of his mother and father. In
the third and fourth grades his response was the same. All the time, however,
musical symbols and objects were appearing on his figural creativity tests. For
example, in third grade, he used the circles to draw the “Beatles,” a drum, a horn,
and the like. Finally, in the fifth grade he excelled so well in a summer music camp
that he had the courage to turn away from the “games” others expected him to play
and for the first time began considering a career in music. Some, of course, never
reach this point, and the consequences are tragic.
Surprisingly, about half of my subjects were consistent in their choices and have
persisted in careers consistent with the future career images they expressed as
children. Using the five different sets of criteria already described in this chapter,
having childhood future career images that they could stick with proved to be a
statistically significant predictor of adult creative achievement. In fact, this indi¬
cator (having or not having future career images that they were in love with) was a
better predictor of adult creative achievement than were indices of scholastic prom¬
ise (IQ).
Since I reached the conclusion that the essence of the creative person is being in
love with what one is doing, I have had a growing awareness that this characteristic
makes possible all the other personality characteristics of the creative person: cour¬
age, independence of thought and judgment, honesty, perseverance, curiosity, will¬
ingness to take risks, and the like. For example, Roger Tory Peterson (Beebe, 1986)
at a very young age was in love with nature, especially birds. He was a social
outcast in his youth because of this. He still will not say what nicknames the
taunting youngsters called him. “It wasn’t very nice,” he says. This has been a
common experience of many of our most eminent inventors, scientists, artists,
musicians, writers, and so on. To maintain an intense love for something and
survive these kinds of pressures, one has to develop courage, independence, per¬
severance, and the like. On the basis of these and other insights, I developed a set of
guides (Torrance, 1983) for creative youngsters that Morgan Henderson and Jack
Presbury called a “Manifesto” and made it into a poster that carries these words:

‘ ‘How to Grow Up Creatively Gifted’ ’

E. Paul Torrance
1. Don’t be afraid to “fall in love with” something and pursue it with intensity. (You
will do best what you like to do most.)
2. Know, understand, take pride in, practice, develop, use, exploit, and enjoy your
greatest strengths.
3. Learn to free yourself from the expectations of others and to walk away from the
games they try to impose upon you.
4. Free yourself to “play your own game” in such a way as to make good use of your
gifts.
Creativity as manifest in testing 69

5. Find a great teacher or mentor who will help you.

6. Don’t waste a lot of expensive, unproductive energy trying to be well-rounded.
(Don’t try to do everything; do what you can do well and what you love.)
7. Learn the skills of interdependence. (Learn to depend upon one another, giving
freely of your greatest strengths and most intense loves.)

Factors in groups

The focus in group research usually has been on decision making. However, in most
instances what makes the decision making worth studying is the creativity that takes
place in the process. I have engaged in at least four programs of research that
investigated the group factors that facilitate or interfere with the creative process.
Each of them involved a series of rather complex experiments and tests, and so it is
possible to deal only with the highlights.
The first set of such experiments was conducted at the U.S. Air Force Advanced
Survival Training School and involved both laboratory experiments and realistically
simulated situations (Torrance, 1955). From 7 years of research and dozens of
experiments and observational studies, I think that the most important finding
affecting creativity in groups was that willingness to tolerate disagreement resulted
in better decisions, higher creativity, and better combat performance. This finding
was cross-validated by several experiments, observational studies, and combat stud¬
ies in Korea (Torrance, 1957b). There were dozens of findings of factors that
facilitated or inhibited the process, but they all went back to the question whether
the factor identified facilitated or inhibited the group’s willingness to disagree, to
consider a wider range of alternatives, to listen to all members of the crew, and to
introduce unusual solutions.
We also conducted a series of leadership studies (Torrance, 1957c). From these
studies, the creative leader seems to be one who, in emergency and extreme
conditions,

1. rapidly restructures the group and the situation,

2. uses the expertness (both air force and civilian) within the crew to get out of
trouble,
3. is willing to take risks (trying new skills, trying new foods, etc.),
4. is willing to make decisions and take action, distributing responsibility, taking
responsibility for decisions,
5. acts outside of his authority, taking care of his men, taking emergency actions not
anticipated by standard operating procedures,
6. is willing to share the dangers and discomforts of the situation, sharing unpleasant
tasks, being able “to take it” despite handicaps, and
7. develops a pattern of mutual support.

My second set of group dynamic studies involving creativity was with elementary
school children in grades 3-6. Even by the third grade, children have adopted many
of the strategies used by adults to limit the creative production of the group (Tor¬
rance, 1963, 1965). Among the strategies observed in the 5-person groups in one
sixth grade class were these;
70 E. P. Torrance

1. Assigning the most creative member to lead the group.

2. Assigning the most creative member to be the recorder.
3. The most creative member (a girl) began with a burst of creativity, but, sensing
disapproval, she became the group clown.
4. The most creative member worked in isolation.
5. The most creative member became counteraggressive when he received pressure
from his teammates.
6. The most creative member was silent and preoccupied, finally leaving the group.

Some excellent strategies were also exhibited. In fact, the most frequently used
strategy was initiating structure, leading, and coordinating. The second most fre¬
quently employed was not leading at all, but filling in the gaps, offering ideas when
no one else would do so.
It was in groups heterogeneously arranged according to creativity scores that
stressful situations arose. In homogeneously arranged groups, a cooperative atmo¬
sphere and positive strategies prevailed. They were less active than the hetero¬
geneous groups. In the groups arranged heterogeneously according to IQ, the group
pressure was on the most intelligent member to produce ideas. Many of them, not
being able to excel at this “game,” were embarrassed that they were unable to meet
the group’s expectations. The homogeneous groups were more creatively produc¬
tive than the heterogeneous groups (Torrance, 1965).
The subjects of the third set of experiments were graduate students arranged in 5-
person groups designed to test specific parts of some group dynamic theories (field,
T group, psychoanalytic, sociodramatic, creative problem solving, and synectics
group theories). The limitations of this chapter make it impossible to discuss the
creativity measures, the treatment procedures, and the theories involved. I shall
only enumerate illustrative findings (Torrance, 1972d).
1. Groups whose members were acquainted with each other produced more original
responses than the unacquainted groups.
2. After working on a problem in a group, “Aha” experiences were reported 5 days
later as occurring during the process (32%), during feedback in class (4%), soon
afterward, the same day (44%), delayed (on the second, third, or fourth day)
(56%), and when they were given the follow-up questionnaire (14%). Fifty-four
percent said that their new insights (ideas, “Ahas,” etc.) occurred through group
interaction.
3. Practice of a systematic, disciplined, creative problem-solving process, after 3 hr of
training, brought few complaints from “freedom-oriented” groups, as determined
by the Runner Studies of Attitude Pattern (Runner & Runner, 1965), whereas the
“control-oriented” groups had significant numbers of such complaints (26% vs.
83%).
4. Freedom-oriented groups produced more original responses than control-oriented
groups, but following an evaluative feedback the controlled groups made more
gains in originality than freedom-oriented groups. Under creative feedback condi¬
tions, however, freedom-oriented groups made greater gains in originality.

The fourth set of experiments on group creativity involved preschool children.

Again, the creativity tests involved in these experiments, the treatment procedures,
and the results are far too varied and too complex to present here. The most
Creativity as manifest in testing 71

important conclusion about the nature of creativity, in my opinion, was that so¬
cialization could be achieved without sacrificing creativity. In my opinion, the two
can reinforce one another. By becoming more socialized, a child can become more
creative, and vice versa (Torrance, 1970).
In one series of studies, I investigated the role of dyadic interaction with 5-year-
olds in facilitating creativity. Twenty-four children were tested alone, and 22 of
their randomly selected classmates were tested in dyads. Dyadic interaction facili¬
tated creative output, and the children persisted longer and seemed to have a more
enjoyable experience.
Creative people are willing to attempt the difficult, and children, too, in order to
learn and grow creatively, must be willing to tackle difficult tasks. To do this,
teachers must create social and peer conditions conducive to attempting tasks of
appropriate difficulty. In another experiment, I hypothesized that 5-year-olds placed
in dyads would be more willing to attempt a difficult task than alone or in class. A
target game (a wastepaper can and bean bags at varying distances) was used in this
experiment. There were 22 children in each of three conditions. The results clearly
supported the hypothesis that 5-year-old children are most willing to attempt a
difficult task in dyads and least willing when performing before the entire class.
The ability to support curiosity by skill in asking questions is important for both
socialization and creativity. In one experiment, I examined the effects of group size
on the question-asking behavior of eight classes in preprimary education. In the test
task, groups were asked to produce as many questions as they could in a 10-min
period concerning a picture. It was required that questions be about things that could
not be ascertained by looking at the picture. Groups of 24, 12, 6, and 4 were
studied. Production was significantly higher in the small groups, although in most
respects 4-child groups seemed to have no advantage over 6-child groups.
Many early childhood educators argue that any imposed structure discourages
creativity, whereas others argue for its necessity. In one experiment with 5-year-old
children, an attempt was made to improve the small-group behavior of children by
increasing the task structure. The children were first asked to draw dream castles
and then decide which member’s dream castle to construct, using Lego blocks. In
the second experiment, an attempt was made to improve this type of behavior
through increasing the group structure by designating one member as leader. In the
first experiment, there were twelve 6-child groups in each condition; in the second
experiment," there were six 6-child groups in each condition. In the first experiment,
the task structure resulted in increases in planning and cooperating behavior, de¬
creases in verbal and physical aggressiveness, and an increase in the judged
creativity of the products produced. In the second experimental manipulation, the
structure increased the planning behavior and the judged creativity of the products,
but failed to decrease the verbal and physical aggressiveness.
In the socialization process, children find increasingly that they are forbidden to
touch and to manipulate objects. In an earlier study (Torrance, 1963) I had found
that the degree of manipulation of the toys used in the product-improvement task
72 E. P. Torrance

was highly related to fluency and originality scores. This suggested the hypothesis
that provision of opportunities to manipulate stimulus objects would facilitate ques¬
tion asking among 5-year-old children. The 48 subjects were randomly assigned
within the classrooms to eight 6-child groups. The manipulation and nonmanipula¬
tion were presented in an alternate-form design in the two classrooms. The stimulus
objects were a plastic bee that made a buzzing sound when twirled and a toy musical
instrument that made a variety of sounds when manipulated. The manipulation
condition, when compared with the nonmanipulation condition, produced a statis¬
tically significantly larger number of questions and larger numbers of hypothesis-
stating questions and questions about puzzling phenomena. The results suggest that
previous findings concerning the facilitating effects of manipulation in the produc¬
tion of inventive ideas can be generalized to question-asking and hypothesis-making
skills.
These are only examples of how testing children may add to our understanding of
the nature of creativity.

Limitations

My understanding of the nature of creativity as manifest by testing creativity is

limited because my efforts to assess creativity have thus far been limited largely to
the rational-thinking view of creativity. Little attention has been given to assess¬
ment of the creativity that goes beyond the rational - the suprarational - not
contrary to reason, but outside the province of reason. I have made some tentative
forays into this domain, but my measures are not well enough developed to include
in this chapter. Of course, my “artistic” definition and the work that extends it are
in this direction. I have proposed that there are some practical abilities included in
this domain and that it is important that we develop means of assessing them
(Torrance & Hall, 1980). I have identified the following such abilities and made
tentative suggestions for measuring them (Torrance & Hall, 1980):
1. The ability to perform “miracles” - the ability to accomplish things that go
beyond logical expectations.
2. Empathy and superawareness of the needs of others.
3. Charisma.
4. The ability to solve collision conflicts for which there are no rational, logical
solutions.
5. A sense of the future.

Summary

I have discussed the nature of creativity as it has been manifested to me by testing

for creativity. Creativity is a multifaceted phenomenon that defies precise defini¬
tion. Many very useful definitions have been offered, but I have chosen three that
have guided my research over the years: a research definition, an artistic definition,
Creativity as manifest in testing 73

and a very simple survival definition. I have sought to translate these definitions
into measurement devices that will then have analogues in learning and in “real
life.” My argument for testing as a legitimate way of learning about the nature of
creativity is based on the fact that test behavior does have analogues in learning
behavior and real life, supported by validity evidence linking test behavior and real-
life creative achievements. Evidence of this link has been presented.
Examples have been given from my research and observations concerning what
has been learned from testing for creativity, as well as what has been learned about
the nature of conditions facilitating and inhibiting creativity, the skills involved in
being creative, the characteristics of the creative person, and group factors that
facilitate and inhibit creative behavior. I admit that these insights are derived only
from tests that represent the rational view of creative behavior. Finally, I urge
investigators to explore assessment of the further reaches of creativity, the suprara-
tional. I identify what some of these abilities are and suggest ways of proceeding.

References

Amabile, T. (1986). The personality of creativity. Creative Living, 75(3), 12-16.

Anderson, H. H. (Ed.). (1959). Creativity and its cultivation. New York: Harper & Row.
Anderson, K. (1964). Creativity is. Unpublished manuscript, University of California, Berkeley.
Bach, R. (1970). Jonathan Livingston Seagull. New York: Macmillan.
Barchillon, J. (1961). Creativity and its inhibition in child prodigies. In Personality dimensions of
creativity. New York: Lincoln Institute of Psychotherapy.
Bartlett, F. (1958). Thinking. New York: Basic Books.
Beebe, M. (1986, June 22). The birdman of Jamestown. Buffalo Magazine, pp. 10-16.
Cochran, W. W. (1955). Elementary patent law. In Training manual for patent examiners. Washington,
DC: U.S. Patent Office, U.S. Department of Commerce.
Crutchfield, R. S. (1962). Conformity and creative thinking. In H. E. Gruber, G. Terrell, &
M. Wertheimer (Eds.), Contemporary approaches to creative thinking (pp. 120-140). New York:
Atherton Press.
Cunnington, B. F., & Torrance, E. P. (1962). Sounds and images. Minneapolis: Bureau of Educational
Research, University of Minnesota.
Cunnington, B. F., & Torrance, E. P. (1965). Sounds and images. Boston: Ginn & Company.
de Bono, E. (1967). New think: The use of lateral thinking in the generation of new ideas. New York:
Basic Books.
de Bono, E. (1975). Think links. Blandford, Dorset, UK: Direct Education Services.
Friedman, F., Raymond, B. A., & Feldhusen, J. F. (1978). The effects of environmental scanning on
creativity. Gifted Child Quarterly, 22, 248-251.
Ginter, E. (1980). Creativity self-experiment. Unpublished manuscript, Department of Educational
Psychology, University of Georgia, Athens.
Goetzmann, K. (1979). Conquering a dragon. Unpublished manuscript, Palmyra, NY.
Gordon, W. J. J. (1961). Synectics. New York: Harper & Row.
Guilford, J. P. (1956). Structure of intellect. Psychological Bulletin, 53, 267-293.
Guilford, J. P. (1959). Personality. New York: McGraw-Hill.
Guilford, J. P. (1960). Basic conceptual problems of the psychology of thinking. Proceedings of the New
York Academy of Sciences, 91, 6-21.
Guilford, J. P. (1986). Creative talents: Their nature, uses and development. Buffalo, NY: Bearly
Limited.
74 E. P. Torrance

Henderson, M., & Presbury, J. (1983). A manifesto for children (Poster). Staunton, VA: Full Circle
Counseling.
Howieson, N. (1981). A longitudinal study of creativity: 1965-1975. Journal of Creative Behavior, 15,
117-135.
Howieson, N. (1984). The prediction of creative achievement from childhood measures: A longitudinal
study in Australia, 1960-1983. Unpublished doctoral dissertation, University of Georgia, Athens.
Khatena, J., & Torrance, E. P. (1973). Thinking creatively with sounds and words. Bensenville, IL:
Scholastic Testing Services.
Kubie, L. S. (1958). The neurotic distortion of the creative process. Lawrence: University of Kansas
Press.
Lefrancois, G. R. (1982). Psychology for teaching: A bear rarely faces the front. Belmont, CA:
Wadsworth.
Martin, N. (1985). Creativity is (Drawings). Unpublished drawings. Athens, GA: Georgia Studies of
Creative Behavior.
Mohan M. (1971). Interaction of physical environment with creativity and intelligence. Unpublished
doctoral dissertation, University of Alberta, Edmonton, Canada.
Osborn, A. F. (1948). Your creative power. New York: Charles Scribner’s Sons.
Pames, S. J. (1962). Can creativity be increased? In S. J. Pames & H. F. Harding (Eds.), A source book
for creative thinking (pp. 185-191). New York: Charles Scribner’s Sons.
Pames, S. J., Noller, R. B., & Biondi, A. M. (1971). Guide to creative action. New York: Charles
Scribner’s Sons.
Patrick, C. (1955). What is creative thinking? New York: Philosophical Library.
Prince, G. M. (1970). The practice of creativity. New York: Harper & Row.
Ribot, T. (1906). Essays on the creative imagination. London: Routledge & Kegan Paul.
Runner, K., & Runner, H. (1965). Manual of interpretation for the interview form III of the Runner
Studies of Attitude Patterns. Golden, CO: Runner Associates.
Selye, H. (1962). The gift for basic research. In G. Z. F. Bereday & J. A. Lauwerys (Eds.), The gifted
child: The yearbook of education (pp. 339-408). New York: Harcourt, Brace & World.
Spearman, C. E. (1930). Creative mind. Cambridge University Press.
Starkweather, E. K. (1976). Creativity research instruments designed for use with preschool children. In
A. M. Biondi & S. J. Pames (Eds.), Assessing creative growth: The tests - book I (pp. 79-90).
Buffalo, NY: Creative Education Foundation.
Stein, M. I. (1953). Creativity and culture. Journal of Psychology, 36, 311-322.
Stewart, G. W. (1950). Can productive thinking be taught? Journal of Higher Education, 21, 411-414.
Taylor, I. A. (1959). The nature of the creative process. In P. Smith (Ed.), Creativity (pp. 51-82). New
York: Hastings House.
Thurstone, L. L. (1952). Creative talent. In L. L. Thurstone (Ed.), Applications of psychology (pp. 18-
37). New York: Harper & Row.
Torrance, E. P. (1955). Techniques for studying individual and group adaptation in emergencies and
extreme conditions. In Air Force human engineering personnel and training research (pp. 286-
297). Washington, DC: National Academy of Sciences/National Research Council.
Torrance, E. P. (1957a). Psychology of survival. Unpublished manuscript, Air Force Personnel Re¬
search Center, Lackland Air Force Base, TX.
Torrance, E. P. (1957b). Group decision-making and disagreement. Social Forces, 35, 314-318.
Torrance, E. P. (1957c). Leadership in the survival of small isolated groups. In Symposium on preventive
and social psychiatry (pp. 309-327). Washington, DC: Walter Reed Army Institute of Re-
search/NRC.
Torrance, E. P. (1963). Education and the creative potential. Minneapolis: University of Minnesota
Press.
Torrance, E. P. (1965). Rewarding creative behavior. Englewood Cliffs, NJ: Prentice-Hall.
Torrance, E. P. (1966). The Torrance Tests of Creative Thinking: Technical-norms manual (research
ed.). Princeton, NJ: Personnel Press.
Torrance, E. P. (1970). Achieving socialization without sacrificing creativity. Journal of Creative
Behavior, 4, 183-189.
Creativity as manifest in testing 75

Torrance, E. P. (1972a). Career patterns and peak creative experiences of creative high school students
12 years later. Gifted Child Quarterly, 16, 75-88.
Torrance, E. P. (1972b). Predictive validity of the Torrance Test of Creative Thinking. Journal of
Creative Behavior, 6, 236-252.
Torrance, E. P. (1972c). Can we teach children to think creatively? Journal of Creative Behavior, 6,
114-143.
Torrance, E. P. (1972d). Group dynamics and creativity. In C. W. Taylor (Ed.), Climate for creativity
(pp. 75-96). New York: Pergamon Press.
Torrance, E. P. (1974). The Torrance Tests of Creative Thinking: Technical-norms manual. Bensenville,
IL: Scholastic Testing Services.
Torrance, E. P. (1979). The search for satori and creativity. Buffalo, NY: Bearly Limited.
Torrance, E. P. (1981). Predicting the creativity of elementary school children (1958-1980) - and the
teacher who made a “difference.” Gifted Child Quarterly, 25, 55-62.
Torrance, E. P. (1983). The importance of falling in love with “something.” Creative Child and Adult
Quarterly, 8, 72-78.
Torrance, E. P. (1984). Sounds and images productions of elementary school pupils as predictors of the
creative achievements of young adults. Creative Child and Adult Quarterly, 7, 8-14.
Torrance, E. P. (1987). The blazing drive: The creative personality. Buffalo, NY: Bearly Limited.
Torrance, E. P., & Arsan, K. (1963). Experimental studies of homogeneous and heterogeneous groups
for creative scientific tasks. In W. W. Charters, Jr., & N. L. Gage (Eds.), Readings in the social
psychology of education (pp. 133-140). Boston: Allyn & Bacon.
Torrance, E. P., & Ball, O. E. (1984). Torrance Tests of Creative Thinking: Streamlined (revised)
manual, Figural A and B. Bensenville, IL: Scholastic Testing Services.
Torrance, E. P., & Hall, L. K. (1980). Assessing the further reaches of creative potential. Journal of
Creative Behavior, 14, 1-19.
Wallach, M. A., & Wing, C. W., Jr. (1969). The talented student. New York: Holt, Rinehart &
Winston.
Wallas, G. (1926). The art of thought. New York: Harcourt, Brace & World.
Wechsler, S. (1982). Identifying the creative strength in the responses to the verbal forms of the Torrance
Tests of Creative Thinking. Doctoral dissertation, University of Georgia, 1981. Dissertation Ab¬
stracts International, 43, 3521A (University Microfilms No. 82-01588).
Wilson, R. C. (1956). The program for gifted children in the Portland, Oregon, schools. In C. W. Taylor
(Ed.), The 1955 University of Utah research conference on the identification of creative scientific
talent (pp. 14-22). Salt Lake City: University of Utah Press.
3 Putting creativity to work

Frank Barron

Is applied, creativity a barbarous notion? What would an artist say, or a poet? Shall
we have applied poetry next? And are we not all of us artists at heart, in our lives? Is
not creativity a spark of the divine in us? Do we not reduce it, reduce ourselves, in
fact, when we subordinate creativity to utility, when we make our creativity serve
extrinsic ends? Is not this call, the call to create, nothing less than a call to give
ourselves to the sacred in human life? Is it not the master, are we not its servitors?
Put creativity to work, indeed!
With all due respect for that archetypically artistic point of view, in this chapter I
urge that in fact we have reached such a point of development of our knowledge of
creativity that it is ready for application and, further, that as a species we are at such
a point of historical no return that it behooves us to muster all our powers, in the
name of the Creator if necessary, to meet consciously the problems that our own
evolution has set for us. So, the thesis of this chapter: Put creativity to work!

The research roots

It has become traditional in historical reviews of psychological research on creativ¬

ity to cite J. P. Guilford’s 1950 presidential address to the American Psychological
Association as a sort of watershed, the beginning of serious empirical work on the
topic. And 1969 is often cited as marking a decline in interest and activity. Actu¬
ally, of course, as one might expect, the development began earlier, was more
continuous and sustained, and has simply moved into a phase in which applications
and evaluation are receiving more attention.
Still, there is some justice in the observation that 1950 marked a turning point.
Earlier research usually had been directed to questions about originality and imag¬
ination. “Creativity” became a popular term that included a lot of earlier work
identified under other names. As late as the mid-1950s, when Scientific American
decided to publish a special issue on “Innovation in Science,” imagination was still
the more commonly recognized rubric. My own contribution to that issue was
written to the title “The Psychology of Imagination” (Barron, 1958). Imagination
somehow belonged to general psychology and was seen as a natural and universal
function of mind, as well known to inner experience as to outer expression.
76
Putting creativity to work 77

Creativity is in fact more of a social outcome, marked by certain kinds of behavior

and having utilitarian value.
There is also some justice in the observation that interest in creativity began to
wane as the decade of the 1960s drew to a close. But again, mass behavior and
social values were the main factors in this waxing and waning. There had been a
sort of self-conscious creativity in the way that social innovations and changes in
style were embraced in the psychedelic sixties, and when it proved hollow, or at
least not yet sound enough to endure in the face of practical problems, creativity
itself got a bad name.
Meanwhile, however, a solid body of research, ready for a less obtrusive but
pragmatically important role, continued to accumulate. The work in the 1950s and
early 1960s had established a beachhead that would be ready for a breakout at a
propitious time. I think the image appropriate; creativity was not at all unlike an
invasion, resisted, or at least not altogether welcome, in conventional, mainstream
psychology.
A look backward at that period of creative ferment seems very much in order,
however, now that we can say with some assurance that it did lead on to a well-
established body of knowledge and to excellent techniques of measurement and
identification. That disciplined knowledge remains somewhat underused, as a mat¬
ter of fact, considering the impressive validity of some of the measures and meth¬
ods. But there has now been a resurgence of social interest in creativity, especially
in the applied aspects of selection and training, as well as in idea production. This
renewed interest is far from faddish, and it is not simply a late blooming that can be
expected soon to fade. The first rapid growth of research got its impetus from the
situation of nation and world in those wide-open days that followed World War II,
when anything seemed possible to the intelligence and the technological aptitudes of
humanity. This second growth comes from the sudden awareness, finally, that
reality has indeed changed, that the problems of human beings are now threatening
the well-being of the planet, and that the species itself is in really serious danger of
extinction. There is no other way to read the world situation as we approach the end
of the twentieth century. Creativity is needed more than ever before. New adapta¬
tions are necessary. The matter is urgent. If psychology can be of use, especially in
the ways in which it has proved itself most conspicuously - measurement of
abilities, identification of talent, selection for challenging assignments, intensive
training, behavior change, and encouragement of new ways of thinking - then it
should be at the ready and should make its readiness known.

The ingredients of creativity

Creativity as an applied discipline began for me in the decade of the 1980s when a
vice-president of Chevron USA called me at my office to ask if I could “spotlight
the ingredients of creativity.” The image of a huge tank brimful of creativity, each
element distinctly colored, leaped to my mind and stayed there as I heard the rest of
78 F. Barron

the story: that Chevron was putting together a very large museum exhibit on the
theme “Creativity: The Human Resource,” and it was to be organized around a list
of ingredients. I allowed as how I could oblige, and we soon agreed on a time and
place for getting it on the road.
In the ensuing months I did literally hit the road from Santa Cruz every Saturday
to get together in San Francisco with Bruce Burdick, head of the well-known
environmental design company, the Burdick Group, which was in fact responsible
for the entire exhibit. (I was chief psychological consultant, but I would not want to
be seen as claiming more credit than is my due; this was Bruce Burdick’s show, and
Chevron’s.)The ingredients, as we worked them out, were the following:
1. Recognizing patterns
2. Making connections
3. Taking risks
4. Challenging assumptions
5. Taking advantage of chance
6. Seeing in new ways

As it turned out, the exhibit was huge, weighing 60 tons when crated for ship¬
ment. We incorporated in it some of the psychological tests of metaphorical ability
and contingent thinking I shall discuss here, putting them into an interactive com¬
puter terminal format that allowed visitors to play at the tests. We also developed a
series of lessons in creativity, for use by schools, packaged with a film strip on the
topic “Creativity, The Human Resource” (Elliot Eisner, professor of art and educa¬
tion at Stanford, was my coproducer on this). Some 12,000 copies of the film and
instructor’s manual were distributed free on request to schools in the urban areas in
which the exhibit was shown.
Finally, one-third of the exhibit was organized around 18 creative careers of
contemporary Americans. Documents and artifacts, supported by many pho¬
tographs and videotaped interviews of the creative people whose achievements were
thus celebrated, were organized in display units so that viewers could see contained
within a small space the historical record of these achievements and the persons
responsible for them.
In brief, the exhibit “Creativity: The Human Resource” put on display a co¬
herent set of ideas, images, words, and persons as instruments of creativity. And
this indeed is how creativity works: through a motivated, purposeful, acting self
using imagination, reason, and the astonishing human ability to put experience into
words. The exhibit traveled the nation as planned, entertaining and instructing more
than 5 million people in 21 major cities, and it can be counted a great success. At
this point down the road, “Creativity: The Human Resource” is parked perma¬
nently in the Pacific Science Center in Seattle, where it continues to be viewed by
large numbers of visitors each year.
In spite of the joshing I took from my fellow academics, as well as from the
Chevron USA people, who never failed to refer to me as “professor” and to make
some remarks about the ivory tower when we met, I felt this to be a very worthwhile
Putting creativity to work 79

use of the knowledge I had gained from the preceding three decades of my research
on creativity. As for the ivory tower, there was some justice in the jibes. The Social
Sciences Building, now Clark Kerr Hall, at the University of California, Santa
Cruz, cannot be seen from anywhere on campus until one is at the door. It is tucked
safely into a grove of redwood trees, and the Performing Arts Building shelters it
from passersby on the main road. Fortunately, there is some research on mapping
going on on the fourth floor, in the cognitive psychology section, and indeed one
needs a very good map to get to Social Sciences. Without belaboring the point, I
think our tower in the redwoods an excellent metaphor for the social sciences vis-a-
vis public policy, a virtual paradigm of the times.

A consensus about creativity

Although I had been amenable to the notion of analyzing creativity into its ingre¬
dients, I could be so easy about it because a lot of defining and spotlighting had
already gone on. Creativity by 1980 had at last merited a 10-year review in the
Annual Review of Psychology (Barron & Harrington, 1981). A full reading of
research on the topic from 1970 to 1980 gave an impression of sustained activity in
empirical psychological approaches to what by now was a field of psychology in
itself. Consider these topics that had emerged as dominant: creativity in women;
hemispheric specialization opposing right brain to left as the source of intuition,
metaphor, and imagery; the contribution of altered states of consciousness to cre¬
ative thinking; new methods of analysis of biographical material and a vigorous new
approach to the study of genius and creativity by way of generational analysis and
historiometry; the relationship of thought disorder to originality; the inheritance of
intellectual and personal traits important to creativity; the enhancement of creativity
by training; creativity and politics; creative environments. In the years from 1970 to
1980, publications in the field had been fairly constant in number, averaging about
250 per year. This variety naturally cost something in the way of uniformity, but in
all of it a core of shared premises could be found. Creativity is central to all of
psychology; it is general psychology, and it reaches out to philosophy, the arts and
humanities, the natural sciences, and human affairs generally in its ramifications.
I myself had authored the first Encyclopaedia Britannica article on the subject
(Barron, 1963a), and I was aware even then of a substantial consensus among
scholars. When I reviewed the field again for an article in the Encyclopedia Ameri¬
cana some 10 years later (Barron, 1971), difficulties of interpretation of empirical
results had become manifest, but the conceptual consensus was, if anything, strong¬
er. The increasing appearance of contradictory findings noted in the most recent
review, the Social Sciences Encyclopedia article on creativity (Barron & Har¬
rington, 1985), was due at least in part to failures of cross-validation when unrelia¬
ble measures and haphazardly chosen samples had been used by investigators who
had only a sketchy knowledge of the now voluminous literature. It was due also,
however, to an increasing maturity that has led to finer interpretations and ques-
80 F. Barron

tions. In spite of this appearance of diffusion, certain points of agreement are found
again and again in journal articles, unpublished theses, and reviews of the literature:
1. Creativity is an ability to respond adaptively to the needs for new approaches and
new products. It is essentially the ability to bring something new into existence
purposefully, though the process may have unconscious, or subliminally con¬
scious, as well as fully conscious components. Novel adaptation is seen to be in the
service of increased flexibility and increased power to grow and/or to survive.
2. The “something new” is usually a product resulting from a process initiated by a
person. These are therefore the three modes in which creativity may most easily be
studied: as product, as process, as person.
3. The defining properties of these new products, processes, and persons are their
originality, their aptness, their validity, their adequacy in meeting a need, and a
rather subtle additional property that may be called, simply, fitness - esthetic
fitness, ecological fitness, optimum form, being “right” as well as original at the
moment. The emphasis is on whatever is fresh, novel, unusual, ingenious, clever,
and apt.
4. Such creative products are quite various, of course: a novel solution to a problem in
mathematics; an invention; the discovery of a new chemical process; the composi¬
tion of a piece of music, or a poem, or a painting; the forming of a new philosoph¬
ical or religious system; an innovation in law; a fresh way of thinking about social
problems; a breakthrough in ways of treating or preventing a disease; the devising
of new ways of controlling the minds of others; the invention of mighty new
armaments, both of offense and defense; new ways of taxing the citizens of a
country by its government; or simply a change in manners for multitudes of people
in a specific period of time (a decade, for example, or a century, or, nowadays, half
a decade, for things seem to be speeding up as we approach the millennium).

To these four points I would add these notes of qualification:

5. Many products are processes, and many processes are products. And a person is
both a product and a process. Each is in a sense “a field within a field” - a field
that never closes, for we are talking about open systems, mutually interdependent,
with no hard and fast line dividing product from process from person.
6. Creativity is not just an ability, but a characteristic of evolving systems. Psychol¬
ogy, by asking how and under what conditions novelty appears in human psycho¬
logical functioning, links itself to the general scientific enterprise of describing the
evolution of forms in the natural world. The Greeks spoke of cosmogenesis, the
cosmos itself being a system, or an ordered, all-inclusive whole, whose origins
were steeped in mystery and lost in time. From chaos came cosmos; out of an
infinite, undifferentiated darkness emerged material forms, which evolved continu¬
ously. (Furthermore, if one pushes chaos hard, one is likely to find a cosmos in it.)
7. Darkness and light, chaos and order, the unlimited and the limited; these are the
terms in which the physical and natural creation were to be understood. And for
each of these there is an analogous psychological state. Darkness is akin to the
unconscious, light to consciousness.
8. Psychic creation, including the creation of the self, is a form of evolution. It is part
of the evolution of consciousness. A new kind of person is potentially a step in
evolution.
9. Psychic creation is also an analogue to sexual procreation. Its stages are conception
(in a prepared matrix), gestation (through time, in intricately coordinated forms and
sequences), parturition (the suffering to be bom, the emergence to light of the
new), and - let us not forget! - the bringing up of baby (a further period of
realization and responsibility following the birth of the idea or image or new self).
Putting creativity to work 81

10. Creation goes hand in hand with destruction. De-structuration often must precede
the building of new structures. The strife between opposites is an important source
of energy for an evolving new synthesis. In brief, there is a thorn to the rose: Watch
out! Good, bad, or indifferent, creativity does not ask permission; it simply arrives.
Change is incessant and inescapable, and most mutations are lethal. New needs will
require new ways of looking at things - and often the unfamiliar is distasteful and
upsetting.
11. Finally, and above all (for this is frequently overlooked in the psychology of
creativity), innovations need to be entertained with criticism, wisdom, and respon¬
sibility if they are to serve human purposes and “the relief of man’s estate.” It is
up to society to receive or to reject the gift (or sometimes just the imposed will) of
the innovator. (The “new order” is not always what we need, and it may even take
a lot of effort to defeat!)

The Berkeley studies of creativity, 1950-1970

Much has indeed been learned about creativity since 1950, when its empirical study
by test and experiment may fairly be said to have begun in earnest. Wartime
selection of combat crews for the U.S. Air Force and of men for irregular warfare
assignments by the Office of Strategic Services (OSS) had led to a lot of practical
knowledge about how to pick people for resourcefulness, flexibility, courage, and,
yes, creativity. “Creativity” was not yet a word being used by the military, but in
fact the men and women who had been recruited by the Manhattan Project for the
making of the atomic bomb were chosen above all for their intellectual creativity,
both theoretical and applied.
When the war ended, most of the gentleman-scholars who had rallied to the cause
of fighting the fascist powers returned to the universities that had been the congenial
settings for the development of the ways of thinking and the techniques of assess¬
ment they had applied so effectively in fighting the enemy. But they returned with a
lot of leftover curiosity about just what they had been doing. Psychologists in
particular had theoretical questions about the nature and measurement of creativity,
questions about the validity of the techniques they had been pressed into using, and
curiosity, even eagerness, about new, nonmilitary applications.
Many lines of inquiry began separately, in this or that university. It fell to my lot
to become involved in the work initiated in 1949-50 at the University of California,
Berkeley, under the leadership of Donald W. MacKinnon, who had headed the
main station of the OSS Selection Service. Given the directorship of the newly
established Institute of Personality Assessment and Research (IPAR), MacKinnon
began to recruit for a rather different purpose and to launch a program of research
into “effective functioning,” a term that was made to embrace originality, personal
soundness, and professional success.
The IPAR staff in 1949 was an unlikely assemblage of psychologists, particularly
so in view of its high morale and its strong identity as a unit. It was enthusiastically
pledged to new explorations in psychology; yet it represented perspectives some¬
what alien to one another, even then. There were two psychoanalysts (Erik H.
82 F. Barron

Erikson and R. Nevitt Sanford), two clinical psychologists (Robert E. Elarris and
Harrison G. Gough), an experimental social psychologist grounded in the gestalt
tradition in cognitive psychology (Richard S. Crutchfield), two more or less uncom¬
mitted graduate research assistants (Ronald Taft and myself), and the director of the
institute, Professor MacKinnon, a personologist in the Harvard Psychological
Clinic tradition, but with a strong interest also in the work of Kurt Lewin and “the
life space.”
MacKinnon, as noted, had come to Berkeley from the OSS, where he had played
a key role in selecting men for behind-the-lines operations during World War II.
Erikson was already known internationally as a leading psychoanalytic theorist of
the life cycle, and he had also written one of the first published psychoanalytical
interpretations of Hitler. Sanford was the leading figure in the Berkeley studies of
racial prejudice, especially anti-Semitism, and authoritarianism. And all three had
been intimately involved in the seminal work in American personology, Henry A.
Murray’s Explorations in Personality.
Crutchfield, too, had been prominent in American military psychology. He head¬
ed the U.S. Strategic Bombing Survey at the conclusion of the war, gathering and
making sense of a mass of survey and interview data. Harris was chief psychologist
at Langley Porter Clinic, doing research on stress and its psychosomatic conse¬
quences, as well as on prognosis in psychotherapy. Gough had just completed his
doctoral work at the University of Minnesota, where he had worked closely with
Paul Meehl and Herbert McCloskey on the measurement of sociopolitical attitudes
as aspects of personality. Taft was an Australian, interested naturally enough in
patterns of immigration and in the logic of prediction.
I had come to Berkeley psychology by way of Cambridge University and the
University of Minnesota. I had brought with me the influences of Frederic C.
Bartlett, whose seminar on thinking and remembering I had taken at Cambridge,
and Richard M. Elliott, through a tutorial devoted to the concept of the self in the
psychology of William James. Imagery, words, and schemata were the main topics
in Bartlett’s seminar, as they were in his classic book. Remembering. His seminar
meshed for me with the Clark Lectures, given that year (1946) at Cambridge by the
poet C. Day Lewis under the title The Poetic Image (Lewis, 1946).
The development of measures, experiments, interviews, and theoretical frame¬
works proceeded apace at IPAR. These new measures, experiments, and tech¬
niques, several dozen in all, were used through two decades of research with groups
of subjects chosen for some special reason, usually related to creativity and/or
personal effectiveness. The special reason often included also a declared career
intention, or membership in a recognized profession. Thus, we studied groups as
diverse as doctoral candidates in the natural and social sciences, flying officers in
the military, psychological warfare officers, physicians, mathematicians, writers,
architects, artists, engineers, businessmen, and mountain climbers.
It would be quite impossible for me in this brief space to give even a summary
account of the results of the IPAR studies, which after all have found publication in
Putting creativity to work 83

several hundred professional journal articles and several books. I have elected
therefore to limit myself to the development of measures arising from my own
philosophic, literary, and esthetic interests. This allows me to focus on what I know
best, while yet permitting a reasonable range in considering the more general
aspects of creativity.

Berkeley as laboratory, Cambridge as catalyst

My first publications as a staff member of the Berkeley group dealt mainly with
personality variables measured by verbal questionnaires: basic religious beliefs,
independence of judgment, and a postulated personality dimension that I called
complexity of outlook. This latter variable seemed to cut across the well-known
conservative-liberal dimension in religion, politics, and social issues; I was thus led
to construct a set of scales for inclusion in a comprehensive inventory of personal
philosophy. But I had worked from the outset in developing quite another set of
measures, cognitive in nature, with emphasis on visual images and words. The first
of these was a measure of preferences among visual displays, in which symmetry
and asymmetry, simplicity and complexity, were the chief dimensions. A second
was an attempted simplification, both as to metric and stimulus complexity, of a
scale for the dispositional tendency to give “human movement’’ responses to
inkblots, introducing the concept of threshold and of precisely graded evocative
power in the stimuli themselves. A third set of measures was concentrated on
discernment of patterns within congeries of letters and words - from anagrams to
jumbled sentences (word rearrangement) to compositional tasks in which the re¬
spondent was given a set of nouns, verbs, and adjectives to be woven into a story
(with instructions to use the given words and as few others as possible to achieve
coherence). The aim of all this psychometric effort was to study the transformation
of images, in modes ranging from the literal to the highly symbolic.
Theory, at that point, certainly was taking a back seat to what was simply a
frontal attack, with both barrels blazing, on the obvious fact that the ability to
change things - to produce or even just to allow transformations - is central to the
creative process. New forms do not come from nothing, not for us humans at any
rate; they come from prior forms, through mutations, whether unsought or invited.
In a fundamental sense, there are no theories of creation; there are only accounts of
the development of new forms from earlier forms.
Granting this, one would, of course, like to know about the conditions under
which novelty is most likely to emerge or can be most easily produced. In the
human case, the study of lives, with special attention to periods of high creativity
and to decisive events in the production of important innovations, offered itself as a
difficult but most promising method. At the same time, transformations that could
be produced in the laboratory, if not by precise experiment then at least in a standard
format for observation, should be given special attention.
This was a point of view I had had impressed on me in the Psychological
84 F. Barron

Laboratory at Cambridge, where I had spent the Lent term in 1946 while still in
service with the U.S. Army. The extraordinary vividness with which I still recall
that brief term at Cambridge comes, in part, I am sure, because it succeeded so soon
the war’s ending. The intensity of the experience came, too, from the contrast
between combat zone and quad, the disorder and sometimes near meaninglessness
of the war and the tranquillity and hard-core reasonableness of the university.
Suddenly it had become possible to reflect on events in the round: the actions in
which I had taken part as a combat medic, the savagery of the bombing of cities, the
shock of the concentration camps, and, finally, the birth and use of The Bomb, that
startling announcement by unexampled violence of intellect’s most awesome tri¬
umph. Reality had indeed changed; there was that fact to absorb, to think about.
Creativity had taken on a new meaning. Humanity had a new image in its mind’s
eye: the mushroom cloud of the atomic bomb. It had a new word: Hiroshima.
Image and word, meaning and changes in meaning - these were the very themes
in Bartlett’s seminar. By chance, C. Day Lewis in the Clark Lectures (The Poetic
Image) was focusing on the same themes, and because my interests were as much in
literature as in psychology, I attended both. (It was not really by chance, of course,
because Lewis was wrestling with new images and needed new words to express
new meanings, as we all did.)
One evening, midway in the lecture series, the poet spoke to a point that was to
become a central concern for me in the psychology of creativity. It was, though he
had different words for it, the problem of gestalt transformation in memory. Later I
was to deal with it as a productive thinking ability, symbolic scope, with emphasis
on individual differences. Its elements are both images (in a variety of sensory
modalities) and words. The process requires the recognition of patterns, and the
genesis, spontaneous or sought, of one pattern from another.
Lewis began his lecture, “The Living Image,” with two examples from his own
experience:

A poet, lying half awake one night on the seacoast of North Devon, listened to the sound of a
storm. The sound presently could be separated into three - a diapason roar, quite steady,
which was the wind chiefly but augmented by a continuous growling of the sea; a higher,
hissing note, variable in pitch and volume but no less intermittent, given off by the surf
seething forward and then sucked back down the pebbles; and, behind these, a rhythmical,
not quite regular throbbing - the beat of the waves on the beach. ... He was suddenly
transported to a day in childhood, when he stood on a platform at a London terminus beside
an express engine. The sound this engine made was identical with the storm. There was a
high irregular hiss of steam blowing off, the deep roar of the forced draught in the furnace,
and the same rhythmical throbbing, all blended into one.

Another time, the poet was looking out of his bedroom window in blitzed London. A
searchlight practice was on. The beams swung about the sky, then leaned together like the
framework of a wigwam, and at the apex an aircraft could be seen, silver, moth-like, flying
slowly, found, lost, found again by the searchlights. It was a common enough sight just
then ... but this time the poet saw it differently, as a dramatic paradox; it seemed to him
that candle-beams were desirously searching for the moth. (Lewis, 1946)
Putting creativity to work 85

Lewis had gone on to use these examples as point of departure for an inquiry into
a more specific question. As he put it, “How far can the poet successfully make use
of objects like aeroplanes and engines in metaphor?” But my own imagination had
been attracted by the more general problem: first, the process of transformation of
the presenting image, and then the originality, complexity, and aptness of the
correlative image, whether produced or recalled (recognized, if you will).
This led directly to the first procedure I constructed for the IPAR studies of
creativity. Though it goes under the name Symbolic Equivalence Test, it is not so
much a test as an occasion for observing the process whereby a standard stimulus
image is changed by design into a nonliteral or symbolic image that is recognizably
another version of the original configuration.

The transformation of images

To study the process of transformation of images, I betook myself first to declared

poets. I began by proposing rather complex images, similar to the two examples
from C. Day Lewis. “A twig is floating down a stream, reaches a point in the
flowing water where a log has been thrown across it, spins slowly in the swirls of
water set up by the log, finally gets around the log and floats on.” Or, “a single
white car, parked all by itself at midnight on a black macadam parking lot, under a
flashing neon sign.”
Images like these proved too difficult, or at least unappealing to my subjects.
Then I remembered a synopsis for children of Melville’s Moby Dick, presented
really as a good fishing story (which, after all, it is). The contrast between checkers
and chess came to mind; children of age six can play checkers and have fun at it,
and in fact they usually think they are playing very well; chess is closed to them,
unless they can make checkers out of it. I wanted to construct a test that could be
used through a wide age range and at quite different levels of ability. And so I
discovered the important principle: Keep it simple, and let the respondent make it
complex. In the lingo of test builders, do not put a ceiling on it, and make the floor
as low as possible.
This is an important point in the psychology of creativity. For the creative
solution is not known beforehand, and there is an immense range of possibility for
new developments once we get deep into a problem. Not only are there no “right or
wrong” answers, there really are no “answers” at all, until they have been tested in
someone else’s perception, or by external reality. And the criterion centers on
qualities like “fitness,” “surprise,” “novelty,” “ingenuity,” and even “ade¬
quacy in meeting a need” (for something equivalent, of course; a common enough
demand in life).
In this sense, the Symbolic Equivalence Test had in it the main requirements for
both the creative process and the creative product. Though essentially a laboratory
procedure, it condensed some of the most important practical features of conditions
86 F. Barron

for creativity in life in the large. Think for a moment of these problems in psycho¬
logical theory: (1) the specification of the properties of a pattern in determining
stimulus equivalence and response equivalence for the study of generalization in
learning; (2) the recognition in dreams and in neurotic symptom formation of the
equivalence of images - traumatic events and their consequences, for example - or
the mechanisms of substitution and displacement in dream figures, not to mention
personages in one’s “real” life who are chosen for love, hate, or indifference
simply on the basis of their symbolic equivalence to someone else who had been
there earlier; (3) the changes a scene may undergo in memory - the whole problem
that set gestalt psychology afire, that is, the retention in memory of the transformed
gestalt features rather than the more elementary or component features of the origi¬
nal stimulus pattern; (4) all the many personifications of Fate, the Soul, the Race,
the Homeland, seasons and skies, and passages from birth to death - the images the
Greeks named archetypes.
To give concreteness to some of these transformations, examples drawn from the
Symbolic Equivalence Test procedure itself may be useful. The instructions to the
respondent are as follows:
In this test, you will be asked to think of metaphors, or symbolically equivalent images, for
certain suggested stimulus images. The task can best be made clear by an example.
EXAMPLE:
Suggested stimulus image:
Leaves being blown in the wind.
Possible symbolic equivalents:
A civilian population fleeing chaotically in the face of armed aggression.
Handkerchiefs being tossed about inside an electric dryer.
Chips of wood borne downstream by a swiftly eddying current.

The test example was not chosen with any particular poetic equivalent in mind.
Shelley’s “Ode to the West Wind,” however, is an elegant, sustained example of
“an equivalent.” Recall the opening lines of that great poem, and the images that
run through the five stanzas:
Oh wild West Wind, thou breath of Autumn’s being,
Thou, from whose unseen presence the leaves dead
Are driven, like ghosts from an enchanter fleeing,
Yellow, and black, and pale, and hectic red.
Pestilence-stricken multitudes . . .

What has Shelley done, if we imagine this to be an elaboration of an image of

“leaves being blown in the wind”?
1. He has specified the wind, to make it the wind of autumn, and therefore the dying
of the year.
2. He has made the presence of oncoming death unseen, save in its effects on the
leaves - presumably once green, they have lost the sap of springtime and the
vigorous hue of summer and are now pale, black, yellow, or hectic red, the last
overwhelmed fling of impulse.
3. The leaves are driven like ghosts, fleeing from a tyrannical enchanter; they are
scurrying, eager to get away.
Putting creativity to work 87

4. The leaves are multitudes: frightened exiles, old folks and orphans, all homeless,
driven by the onward rush of pestilence, an unchecked destructive power.

Interestingly enough, the possible answers suggested in the test example do have
the idea in them of a chaos within a coherence (the tossed handkerchiefs, the bereft
victims of the winds, of the pestilence, of war, the speed and the chance eddies of a
current immensely stronger and deeper than the chips of wood; the chips are parti¬
cles, like the leaves, and they are massed yet in disorder and subject to the whims of
the greater natural power).
Shelley goes on to invoke the cycle of the seasons to offer hope:
The winged seeds, where they lie cold and low,
Each like a corpse within its grave, until
Thine azure sister of the Spring shall blow
Her clarion o’er the dreaming earth . . .

And then, rebirth from the ashes and the unextinguished creative spark:
Drive my dead thoughts over the universe,
Like wither’d leaves, to quicken a new birth;
Scatter, as from an unextinguished hearth
Ashes and sparks, my words among mankind.

Poems did come from some of our test stimulation, I might add, and although
Shelley was not there, we did “assess” some famous writers and architects who
have a claim to genius. A test response is not yet a poem, of course; the complex
integral act is a very different matter, as we never forgot.
But back to the laboratory. Following the example, the test respondent is in¬
structed to make up three possible equivalents for each of 10 images. A time limit of
20 min is used for this form of the test. The stimulus images are as follows:
haystacks seen from an airplane; a train going into a tunnel; the sound of a foghorn;
a candle burning low; a ship lost in fog; a floating feather; the increasing loud and
steady sound of a drum; sitting alone in a dark room; empty bookcases; tall trees in
the middle of a field.
The Symbolic Equivalence Test was scored for (1) number of acceptable or
admissible, but not original, responses and (2) number of original responses. Ad¬
missible responses were in turn differentiated as to degree of aptness, scored as 1, 2,
or 3; original responses were differentiated as to degree of originality, scored as 4 or
5. The total score on the test was simply the sum of the ratings of the individual
responses. For purposes of analysis, several additional subscores were defined:
percentage of original responses; absolute number of original responses; absolute
number of responses (admissible plus original).
Here are some examples:

1. Stimulus image:
A candle burning low
Admissible responses:
Life ebbing away (scored 1)
88 F. Barron

A basin of water emptying down a drain (scored 2)

The last drops of coffee going through a filter (scored 3)
The last pages of a faded book (scored 4)
The last hand in a gambler’s last card game (scored 5)
2. Stimulus image:
Empty bookcase
Admissible responses:
A hollow log (scored 1)
An empty sack (scored 2)
An abandoned beehive (scored 3)
An arsenal without weapons (scored 4)
A haunted house (scored 5)
3. Stimulus image:
Sitting alone in a dark room
Admissible responses:
Lying awake at night (scored 1)
An unborn child (scored 2)
A stone under water (scored 3)
A king lying in a coffin (scored 4 or 5)
Milton (scored 5)

Even the lowest ranking of admissible responses moves some distance from the
literal correspondence, as these examples show. To be admissible at all, the re¬
sponse must reproduce the main features of the stimulus image, which included
formal properties as well as functional relationships among them. A candle burning
low, for example, includes prominently (1) a source of light and perhaps heat, (2) a
process that uses up a supply of something, and (3) visible evidence that the process
is coming to a close. An empty bookcase has the characteristics that (1) it is a
container, (2) its earlier or usual contents are many and can be numbered, and (3)
the contents were valuable or useful. Sitting alone in a dark room (1) implies a
personal presence, (2) combines solitariness with darkness, and (3) specifies an
enclosure.
Original responses are, first of all, unusual. Once a catalogue of responses has
been assembled from many subjects, an unusual response is easily recognized. To
be highly original, or to get the highest rating, the response should have additional
dimensions: It grabs you, it surprises you, it gives you a chill as a great line of
poetry can do. And even here there are degrees of elegance. A king in a coffin is
unusual and gripping as an image; a king in his grave might perhaps be better (or
perhaps not); Milton (after his blindness, we are left to understand) is the greatest of
kings in the darkest of rooms.
Needless to say, there are problems in scoring, in this as in all imaginal produc¬
tions. Literal (concrete, physical) complexity must be distinguished from concep¬
tual complexity; the preferences of the raters for complex versus simple phenomenal
fields must be taken into account; the ability of the rater to recognize transforma¬
tions and to differentiate them as to quality must be reckoned with.
Putting creativity to work 89

The Symbol Equivalence Test as a measure of individual differences

in symbolic scope

Eventually the Symbol Equivalence Test was administered to several hundred very
creative people in the IPAR studies of artists, architects, writers, mathematicians,
scientists, and mountain climbers (these mountain climbers, by the way, were the
first American team to ascend Mount Everest, and the first team ever to do it by the
West Ridge - a strenuous way to challenge assumptions, take advantage of chance
and the weather, see things in a new way, take risks, and so forth). The climbers
were “assessed” just before they left San Francisco for Nepal in 1963, accom¬
panied, to be sure, by an IPAR staff member, Dr. James T. Lester, Jr., who made
the climb with them and assiduously got them to tell him their dreams en route.
The validity of the Symbol Equivalence Test as an ability measure proved quite
substantial. First of all, it provided a clear rank ordering by creativity of samples
studied, highly correlated with group rank when they were rated independently for
overall verbal creativity. Famous writers topped all groups, followed by famous
architects, and then in turn by (1) mathematicians at the Institute for Advanced
Study in Princeton, (2) women mathematicians who had been rated as the most
creative in their field in the United States, (3) successful entrepreneurs in Ireland,
(4) the Mount Everest team, (5) research scientists, (6) student artists at the San
Francisco Art Institute, and finally a melange of “opportunity samples” who were
assessed by test only and for a variety of reasons during the 1950s and early 1960s at
IPAR. An important point here is that the test responses were scored by an elabo¬
rate, laborious, and expensive procedure in which every response image was rated
independently by three skilled raters without knowledge of other responses given by
the same subject, and also, of course, without knowledge of the individual’s name
or group membership. (Later, by the way, I used the norms thus established to
develop a mechanically scored multiple-choice form of the test, Symbol Equiv¬
alence II.)
The Symbol Equivalence Test scores were then correlated with external criterion
ratings of creativity of each subject in relation to all others in the relevant profes¬
sional group. Substantial positive, statistically significant correlations were found
for writers, architects, artists, and entrepreneurs. Details of these studies are avail¬
able elsewhere (Barron, 1969).
In brief, a relatively short and simple test, taking half an hour of administration
time and another half hour for a trained scorer to evaluate, has demonstrated
substantial validity for measurement of a key component of creativity: the ability to
make original and apt transformations of a given image, received in words and
expressed in words.
Originality in producing symbol equivalences has been correlated with other
measures as well in several of the samples. Research with student artists provides a
good example (Barron, 1972b). In that study, the subjects were 92 art students in
90 F. Barron

studio classes at the San Francisco Art Institute. They took a battery of psychologi¬
cal tests at the beginning of their first year of studio work. The test battery included
a scale for rating oneself on creativity, an adjective self-description questionnaire
scored on an “artistic self-concept” key, the Barron Independence of Judgment and
the Barron Complexity of Outlook scales, and the following set of performance
measures: the McGurk Perceptual Acuity Test, the Barron M-Threshold Inkblot
series, the Gottschaldt Embedded Figures, the Franck Drawing Completion Test
(Barron, 1958), and two forms of Symbol Equivalence: Form I, free response
(scored for originality) and Form II, multiple choice (scored for “recognition of
originality”).
The two forms of the Symbol Equivalence Test used were SE-I, scored for
overall quality, and SE-II, scored for recognition of originality, a multiple-choice
format presenting two alternative responses to each stimulus image, one of which
had been rated “commonplace” and the other “original” in the earlier normative
study. The instruction to the test respondent on SE-II was to “pick the more original
response.” The overall free-response quality score showed a high positive correla¬
tion with other variables in the test battery and in studio performance were quite
similar: significant positive r’s with perceptual acuity (accuracy of perceptual judg¬
ment and resistance to illusion), ability to recognize a previously seen pattern
embedded in a more complex pattern, quality of drawing on a standardized drawing
completion test, and independence of judgment and complexity of outlook on two
inventory scales, and finally, and most important to the Admissions Office, sub¬
stantial correlations (.60) with studio grades. Grades, in turn, by the middle of the
third year at the Art Institute, were correlated positively in the neighborhood of .50
to .60 with initial self-ratings, complexity of outlook, consequences, and a strong
tendency to perceive human movement in the Barron M-Threshold Inkblot series (as
well as with the scores on SE-I and SE-II).
These findings reflect what is generally considered the core of the metaphoric or
analogical process. They add up to a picture of substantial construct validity as well
as predictive validity for the Symbol Equivalence procedure. But what of transfor¬
mations more dependent on rational imagination, logical processes, contingent
thinking? These, too, are decisively important in creativity, particularly in the
disciplines and professions that make things happen in the great world - science,
technology, law, politics, business, the military.

Transformations in the service of prediction and control

Poets are supposed to be impractical. Planners are not. Contingent thinking, or

discursive reason applied to problems of prediction, is the mark of the practical
individual. These are stereotypes, of course, but like most stereotypes, there is
some truth to them.
Stereotypes become most misleading when they are personified. John Doe is a
Putting creativity to work 91

poet - ergo, impractical. Jane Roe is in business - ergo, bound to dollars and cents,
the bottom line.
There are indeed archetypal poets and archetypal merchants, but we avoid the
danger of stereotypes if we think in terms not of persons but of modes. In creativity,
the analogical mode often is at war with the logical, but the creative act issues from
the tension between the two. One without the other is lost.
Creative people are equally capable of the logical and the analogical, are open
alike to the rational and the nonrational, are at home to image and to abstraction.
This hypothesis argues for another sort of test, a measure of rational thinking in the
domain of action, or at least of anticipation of the consequences of an action or
event.
Several such tests had already been developed elsewhere by 1950, and IPAR put
two of them to use in its assessment program. One was the Bennett Test of Produc¬
tive Thinking. For an example of its use, see the chapter “An Odd Fellow’’ in
Creativity and Psychological Health (Barron, 1963b). It asks questions such as
“What would happen if the mean level of the oceans were to drop 5 feet?” Another
test, developed by J. P. Guilford and his colleagues, was the Consequences test
cited earlier, similar in format to the Bennett test. For example: “What would be
the result of ... if everyone could read everyone else’s mind?” The tests were
scored in slightly different ways, but basically they were made to measure the
ability to produce ideas about the future given certain contingencies.
I developed a scoring scheme for the Bennett test based on the dichotomies of
banality-originality and practicality-cosmicality. It was evident from the outset
that the content of the test (i.e., the specified contingencies) was highly modifiable,
and I soon began using the test with items composed for specific purposes. In our
study of Irish entrepreneurs, for example, I made up a “Consequences for Ireland”
form of the Consequences Test. Working with State of California heads of divisions
in a workshop sponsored by the California Management Institute, I proposed con¬
tingencies relevant to planning for the future in California (water, population,
immigration, the master plan for education, etc.). For a meeting of the Society of
Gynecological Laparoscopists, I gathered data on “what would happen if in vitro
methods for fertilization of human ova became medically safe and inexpensive?”
This latter question is part of my own form of the test, now known as the “What-If
Test,” to elicit people’s thinking about what would happen given certain contingen¬
cies in the “nuclear-space-computer” age, especially concerning nuclear arms
and nuclear energy situations.
The point of all this is that tests like Symbol Equivalences and What-Ifs can be
adapted to virtually any set of problems and can be coordinated with one another.
They are extremely effective as points of departure for workshops and can be used
not only to activate imaginal processes but also to focus practical discussions. By
choosing stimulus images from the contingencies themselves, the analogical mode
can be used to enrich the analytic and predictive mode. I am sufficiently encouraged
92 F. Barron

by results I have obtained in workshops to believe that rigorous experimentation

along this line would pay off. In effect, the construction of symbolic equivalences
for practical terms in a contingent thinking problem is a way of “looking aside”
and of energizing abstract reason by metaphor and nonliteral images.
One example may suffice. In a group workshop setting, I divided the participants
into two working committees. The task of Committee A was to bring in recommen¬
dations on how to spend a given amount of civic funding for a Senior Citizens
Center. The task of Committee B was set by reading to them the opening lines of
Coleridge’s poem “Kubla Khan,” known in the creativity literature because of the
alleged interruption of Coleridge’s reverie by the infamous “traveler from Por-
lock.”
In Xanadu did Kubla Khan
A stately pleasure dome decree
Where Alph the sacred river ran
Through caverns measureless to man
Down to a sunless sea.

The question for Committee B was, “What sort of stately pleasure dome would
you decree?” (The symbolic equivalence is here quite complex, of course, because
the poem hints of last pleasures before descent to the eternal shade.)
This was an unpublished study, but I did count the ideas and evaluate their
originality under two conditions for each subject: serving on Committee A first
versus Committee B first. The prior “thinking-aside” condition employing the
Coleridge poem as a symbolic equivalent led to much greater originality when the
“real contingency” was presented.

Personal philosophy and creativity

Recall that symbolic scope as measured by the Symbol Equivalence Test proved to
be highly correlated in art students with two of the measures I had developed in the
early 1950s as part of the “tooling-up” process for the IPAR studies: (1) indepen¬
dence of judgment and (2) complexity of outlook. These two variables, measured
by inventory scales, are now part of the “Inventory of Personal Philosophy” (an
unpublished values inventory developed for research by the present author and
available from him on request). Both measures were found in early studies to be
substantially predictive of a wide range of characteristics associated with creativity,
including cognitive measures of flexibility as well as social traits indicative of
acceptance of racial group differences (low ethnocentrism), changing mores sexu¬
ally, ecumenical attitudes in matters of religious belief, and internationalism.
The scale for measuring independence of judgment was developed in collabora¬
tion with Solomon Asch and was based on significant item correlations with inde¬
pendent (nonyielding to conformity pressures) behavior in the Asch experiments
(Barron, 1953b). The Complexity of Outlook scale was developed (Barron, 1953a)
by successive item analyses in several samples against preference for complex-
Putting creativity to work 93

asymmetrical versus simple-symmetrical visual displays (the Barron-Welsh Art

Scale). Extensive use of the verbal scale in succeeding samples provided the basis
for naming it Complexity of Outlook.
The two scales rest on a powerful theoretical base with ramifications in scientific
and artistic creativity. To see things as they are, not as the conforming masses
would have one see them, is the hallmark of realistic unconventionality. It requires
accuracy of perception, persistence in seeking the truth, and personal force and
courage in resisting the inevitable pressures to conform. The famous Asch experi¬
ments condensed these requirements into the experimental social situation he
devised.
To prefer a complex phenomenal display is not in itself a mark of creativity,
unless it is coupled with an unrelenting drive to find a simple order. Simplicity in
complexity, unity in variety - these are the criteria for beauty and elegance, in art,
mathematics, science, and personal consciousness in general.
A coupling of these two characteristics, independence of judgment and complex¬
ity of outlook, is potentiated by symbolic scope and by contingent thinking that is
both realistic and searching.
Strong support for this idea is provided by recent results. In a 1983 lecture
(Barron, 1983) I reported correlations between scales of the Inventory of Personal
Philosophy and originality as measured by What-If Test responses to nuclear con¬
tingency items. (Examples: What would happen if the ozone layer were destroyed?
What if terrorists were to obtain possession of a dozen thermonuclear weapons?
What if one country and no others were to develop an impenetrable defense against
ICBMs? What if a limited nuclear war were to begin in the Middle East? Originality
in responding to these questions was correlated positively in the neighborhood
of .50 with both Independence of Judgment and Complexity of Outlook, and nega¬
tively with Ethnocentrism (.35).
An item analysis of the entire inventory in its 1983 revision showed that many
items specific to attitudes about nuclear arms control, scored so as to reflect a belief
that an immediate freeze on nuclear testing and substantial reductions in nuclear
weapons worldwide would be highly desirable, were also correlated highly with
both Independence of Judgment and Complexity of Outlook. The Inventory of
Personal Philosophy (IPP) had been expanded in 1983 to include 45 items related
not only to arms control but also to ecological concerns, space exploration, belief in
the likelihood of extraterrestrial intelligence, and internationalism or “global con¬
sciousness” in general. Twenty-nine of these new items were found to constitute
independent factors in the IPP, and further work established the scale as an impor¬
tant differentiator of attitudes in samples in West Germany as well as in Arizona and
California (Bradley, 1984). A subset of 10 of these 29 items, scored as the Nuclear
Arms Reduction Scale, was later correlated in those and in additional U.S. and
European samples with Independence of Judgment and Complexity of Outlook,
yielding r’s once again ranging around a mean of .50. Because there was no item
overlap, the finding is strongly indicative of a possibility that has been largely
94 F. Barron

overlooked in U.S.-Soviet negotiations: that the true attitudes of negotiators and

the actual determinants of outcomes at the negotiating table are personality vari¬
ables. Political scientists, and even social psychologists, have resisted such in¬
terpretations, but these findings deserve further exploration. It is quite possible that
creativity is not being put to work on this most critical problem simply because
personal traits and tangential philosophical preferences are blocks to creative
thinking.

Symbols for self and career

It is easy to lose sight of the self when one is thinking in terms of test factors, even
when they refer to such dynamically important variables as personal philosophy,
symbolic scope, and contingent thinking. But only the self creates in a sustained
manner. To achieve unity of being requires a self to integrate the schemata, the part-
selves. And the more original, complexly inclusive, and deeply experienced the
self, the greater its power of creativeness.
Note here that I speak of the self, not of “personality.” Creative people come in
all shapes and sizes, all colors and ages, and all sorts of personality types. There are
introverts and extroverts, manics and depressives and manic-depressives, schiz-
othymes and hysterics, sociopaths and good citizens - in brief, a wide variety of
temperaments among people who are especially creative. It is not, in my judgment,
the personality type that makes the difference, but the self. To the self I ascribe
motivation and style, the choice of meaning, and the making of meaning, in work,
career, life course. It is in the self that opposites are in strife, calling for choice.
“A career is a general course of action through life.” This dictionary definition
does ask us to think further about “the life course.” For psychology, a concep¬
tualization of the life course is very important. The course of life is by no means a
straight line, or a series of unbroken circles on the face of the clock. It is not just a
progress through time. Lives and careers are spatiotemporal processes that spill over
in all directions, hodologically, topologically, hologrammatically. They exist as
structures and as ensembles - of other people, of documents, of buildings, of
compositions, of the trees we plant and all the other living things we shelter and
help to grow; offspring and creations of all sorts, including ruins, alas! These
ensembles endure through time and ramify in space; they undergo changes as the
world changes and time passes; and some of their effects may be very distant indeed
from their point of origin. Start a style, write a poem, provide a method, enjoin
others to action - who plants a tree plants a hope - these are creative works not only
in themselves but also in their capacity to create new conditions in which the
creativity of others may find expression. I like the poem Seamus Heaney wrote for
Harvard’s 350th birthday, and because we have been hearing from poets in this
essay, I yield to the temptation to quote:
A spirit moved. John Harvard walked the Yard.
The atom lay unsplit, the West unwon.
The books stood open, and the gates unbarred.
Putting creativity to work 95

These words and images celebrate an idea, a place, a space, time and events
unforeseen, a moral and intellectual career; they give us a hint of the sometimes
brooding quality of creative thought. It is not John Harvard’s personality that would
interest me, but his self.

Self and creative career

A creative career is the sustaining of creativity in a coherent enterprise, such as a

profession or vocation or institution, over a fairly lengthy period of time, from
months to years to a virtual lifetime.
To choose a career that is merciless in its demand for creativity is to make a life-
staking, often lifelong, commitment. Even to consider such a choice is to engage
the deepest resources of the self. This sort of career is intrinsically moral and
intellectual in nature. It is not to be chosen idly. It requires an exacting discipline,
toil without stint, and sacrifices, and it yields an uncertain reward. And in the eyes
of the world, there are losers as well as winners in such careers.
Of course, not all careers need engage the individual in such intense self-develop¬
ment, either as a requirement or as an opportunity. There is a middle ground
between that archetypal artist and the archetypal burgher that is being claimed
increasingly today. People do want their occupations to be self-expressive. And
with increasing mobility and leisure there is more opportunity for us to use our
creative potential ad lib. Further, as work itself changes, part of it may call for
creativity at times, but not all the time. Creativity in work is a potential whose
vitality may endure through long stretches of inactivity and then come into action
when called on. There is no need to be creative as a life work.
One’s life itself is there to create. Family, friends, and associates, personal style,
inner experience, general behavior, dress, character - these are creations in the
opportunity available to us all as human beings.
It is in personal style, as aspect of selfhood, that I find the ingredients of
creativity. Openness to new ways of seeing, intuition, alertness to opportunity, a
liking for complexity as a challenge to find simplicity, independence of judgment
that questions assumptions, willingness to take risks, unconventionality of thought
that allows odd connections to be made, keen attention, and a drive to find pattern
and meaning - these, coupled with the motive and the courage to create, give us a
picture of the creative self.

Practical steps toward applied creativity

The term “brain drain’’ is not fully appreciated. The drain is not from one country
to another - the emigration of talent from less rewarding places of employment to
more rewarding is a small and not very important adjustment. The real drain is the
loss of the creativity that stays at home but is not employed properly. And that is a
worldwide waste: brains down the drain (if one likes that metaphor; I prefer Thomas
96 F. Barron

Grey’s “Elegy in a Country Churchyard,” where a “mute inglorious Milton” lies

interred). The word is waste - a waste of talents.
There are numerous ways in which our resources of creative intelligence are
wasted:
1. Creative potential is not identified systematically and nurtured responsibly.
This failure begins in the school system in the elementary grades, and it continues
right through the undergraduate years at most colleges. Not until graduate school is
there any explicit crediting of creativity as an important qualification for admission
to training in independent intellectual work and social leadership.
There are effective means now available for identifying creative potential. In this
essay I have focused on three measures of what I consider significant characteristics
of creativity: symbolic scope (analogical power), rational imagination (applied to
contingent thinking), and two personality traits: (1) independence of judgment and
(2) a dynamic tension between the striving for simplicity and the entertaining of
complexity. Add to these the motivation to be creative, which is open to education
and responsive to incentives, and we have the means for identifying and nurturing
creativity.
The nurturing of creativity in schools requires, it seems to me, direct instruction
aimed at increasing creative thinking abilities: new curricula, new incentives, new
teacher training.
2. Established organizations in government, industry, and education do not take
creativity as a value and consequently do not make provision for creative use of the
creative individuals they employ.
Consider the creativity of university faculties, as an example close at hand for the
most likely readers of this essay. Routine committee assignments and routine cler¬
ical tasks are uniformly accepted as part of most faculty jobs, and to some extent
this is unavoidable and perhaps as it should be. But surely some other institutional
means can be found, short of sending people on leave to “think tanks” (small in
number), for allowing highly trained academic talent to get together with the ex¬
plicit goal of considering society’s needs for innovation. Interdisciplinary programs
may founder when their purpose is defined solely as interdisciplinary, but if the
purpose is socially transcendent, rather than merely accommodative of inter¬
disciplinary viewpoints, surely something more creative is possible. I would like to
see this tried, not just on university campuses but in other organizations of already
highly trained people whose innovative talents are being wasted simply for lack of
an organizational structure that would facilitate creative thinking. Explicit use of
techniques of rational imagination and the search for symbolic equivalences, as well
as other procedures known from research and development in the area of creativity,
should be given a chance.
3. There is at present no national or international recognition that creativity is
itself a product. We are accustomed to indices such as Gross National Product, the
Consumer Price Index, and so on; but where is there the faintest hint that the
important reality, a nation’s creativity, can be measured, studied, and influenced?
Putting creativity to work 97

The components of such a National Creativity Index would have to be identified,

and the means of measurement established, but the problem surely can be solved.
There are countable evidences of artistic, musical, and literary creativity, of tech¬
nological innovation, of influential changes in government through court decisions,
legislation, and executive initiatives. Medicine is already highly organized for
appraisal of outcomes from new diagnoses and treatments, delivery of health care,
and so on. With effort, we should be able to appraise our level of creativity,
nationally and internationally, and pay heed to the cultivation of this human re¬
source. (Nor should we forget creativity’s other face: “destructivity.”)
4. There are virtually no “centers for research on creativity” in existence,
certainly none at the national level. Perhaps we should be wary of a recommenda¬
tion that there should be. Institutionalization can easily become the burial vault of
creativity, and one would want to watch such centers closely. But certainly there is
a place for concerted new research efforts in creativity as a discipline. There are
untapped resources crying out for research: the creativity of women; use of new
technologies (computers, audiovisual educational methods, new environments for
creativity); the genetics as well as social psychology of creativity; the role of
memory in creation (about which virtually nothing is known); psychotropic drugs
and creativity (a topic on which the books should not be closed); emergent social
forms in the developing nations; creativity in relation to religion and philosophy.
These are merely hints at the topics needing research. Considering the vastness of
this human resource and its crucial importance at this juncture in human develop¬
ment, shamefully little is being done. Where should we begin?
We must first recognize that creativity is a human resource, the gift of life to the
human species. It is a unique force in the universe.
The main ingredients of creativity have by now been identified. What shall we do
with all these ingredients? Put them together properly, and put them to work. There
are rivers to be washed, air to be cleaned, mouths to be fed, guns to get rid of,
diseases old and new to be conquered, justice to be ensured to all, new opportunities
to be created, communication and empathy to be increased, secrets of the cosmos
and of the mind to be understood, a changed reality to be comprehended and
adapted to.
Applied creativity is not a barbarism, so far as I am concerned. The more we
know about creativity, the more use we can make of it for the good of all. It is a gift,
a talent to be exercised in the measure to which we have it - one talent or ten, we
should put it to work.
The transformative power of imagination, coupled with the will to apply it, and
given the means for its cultivation in education, government, and the world econo¬
my, is the main source of hope to counter what may otherwise become a pessimism
about the human future as the millennium comes to an end. The dark side of the
millennium is the possibility that earth, too, may die, even in this generation.
Creativity in the face of this darkness is above all a reason for hope, a symbol of the
renewal of earth, when new growth begins.
98 F. Barron

References
Barron, F. (1953a). Complexity-simplicity as a personality dimension. Journal of Abnormal and Social
Psychology, 48, 163-172.
Barron, F. (1953b). Some personality correlates of independence of judgment. Journal of Personality,
21, 287-297.
Barron, F. (1955). Threshold for the perception of human movement in inkblots. Journal of Consulting
Psychology, 19, 33-38.
Barron, F. (1958). The psychology of imagination. Scientific American, 199, 150-166.
Barron, F. (1963a). Creativity (psychology of). Encyclopaedia Britannica, 6, 711-712. University of
Chicago Press.
Barron, F. (1963b). Creativity and psychological health. New York: D. Van Nostrand.
Barron, F. (1969). Creative person and creative process. New York: Holt, Rinehart & Winston.
Barron, F. (1971). Genius. Encyclopedia Americana, 12, 420. New York: Americana Corp.
Barron, F. (1972a). Towards an ecology of consciousness. Inquiry, 15, 95-113.
Barron, F. (1972b). Artists in the making. New York: Seminar Press.
Barron, F. (1979). The shaping of personality. New York: Harper & Row.
Barron, F. (1983). The influence of personal philosophy on creative thinking in the nuclear gun-pointing
situation. Faculty research lecture, University of California, Santa Cruz (available from Laboratory
for the Study of Lives, UC Santa Cruz).
Barron, F., & Harrington, D. (1981). Creativity, intelligence, and personality. In Rosenzweig, M. R., &
Porter, L. W. (Eds.), Annual Review of Psychology (Vol. 32, pp. 439-476). Palo Alto, CA: Annual
Reviews.
Barron, F., & Harrington, D. (1985). Creativity. Social Sciences Encyclopedia (pp. 167-169). London:
Routledge & Kegan Paul.
Bartlett, F. C. (1932). Remembering. Cambridge University Press.
Bradley, P. (1984). Dimensions of personality and attitudes towards the nuclear situation: A study of
West German values. Master's Abstracts (University Microfilms No. 39791/1984/361).
Chambers, J. A., Sprecher, J. W., & Barron, F. (1980). Identifying gifted Mexican-American students.
Gifted Child Quarterly, 24, 123-129.
Lewis, C. D. (1946). The poetic image. London: Jonathan Cape.
Williams, W. C. (1949). Pictures from Breughel. New York: New Directions.
4 Various approaches to and definitions of
creativity

Calvin W. Taylor

Creativity is “a many splendored thing,” as Guilford has pointed out repeatedly.

To us, as displayed in our own studies, creativity is a very complex human perfor¬
mance and occurrence, one of the highest-level performances and accomplishments
to which humankind can aspire. To many in the arts, including poets and creative
writers, the highest degree of the creative process is almost a combined total-
human-being response, involving all aspects of such a person’s response repertoire.
Many subcomponents of creativity are combined simultaneously and/or suc¬
cessively to approach this almost total-human-being response.
Our initial focus and opportunity to undertake major studies on creativity was
among scientists and engineers in large research centers. These studies measured
products, performances, and other accomplishments in their careers, plus their
needs to have better organizational climates for creative scientific work. As we
studied the processes scientists used in their work, we contemplated what is meant
by ‘‘the scientific method.” To us, the colorful and best description is that creative
scientists use the approach of ‘‘no holds barred” to tackle the challenges in an
existing field or to open a new field for scientific investigation. The best scientists
seem to get extremely involved in a problem. They become so immersed that they
stick to the problem, and the problem sticks to them. One fellow scientist described
his colleague by saying that the only way one could stop him from working on his
problem would be to shoot him. All this led to the book that summarized our first
three creativity research conferences: Scientific Creativity: Its Recognition and De¬
velopment (Taylor & Barron, 1963).
A few years after completing a dissertation on fluency in ideas (ideational fluen¬
cy) and on expressional versatility (or fluency), I was invited by the National
Science Foundation (NSF) to take a 2-year leave of absence to work on NSF’s
challenge of selecting the best scientists both for the future of science and for the
good of the nation. By then, fellowship support was provided for predoctoral and
postdoctoral work through the NSF’s new Fellowship Program. My base was the
Office of Scientific Personnel’s first director of research in the National Academy of
Science-National Research Council (NAS-NRC). Through its past fellowship awar¬
dees, the NRC Fellowship Program had already attained a distinguished reputation.
As I left to return to Utah, the leaders and advisors of both NSF and NAS-NRC
99
100 C. W. Taylor

gave me the time-unlimited challenge of studying and learning how to identify

highly creative scientists, a few of whom can make such a tremendous difference in
a field, as had occurred, for example, in the splitting of the atom and the opening of
the Atomic Age. This challenge also included stimulation of others to work in this
realm by holding periodic and timely research conferences among researchers who
had done the best studies on creative scientists. Eventually this led to Widening
Horizons in Creativity (Taylor, 1964), the volume from our Fifth Creativity Re¬
search Conference, with its coverage being expanded to include all areas of human
endeavor in which creativity can function and flourish.
A few comparisons of a highly general nature will be made between the early
Intelligence movement (including its current outcomes) and the early Creativity
research-and-implementation movement, as well as its recent and possible future
outcomes. A concern here is that the Creativity research-based movement definitely
should not follow all the same paths and outcomes of the Intelligence movement
efforts. It appears, in hindsight, that the potentially vast concept of Intelligence
converged and comparatively rapidly was reduced to a much smaller concept, as
measured by group-tested IQ scores (already obtained for probably over a hundred
million persons).
The much earlier Intelligence movement rushed prematurely and noncreatively to
considerably simplified, functional solutions. It materialized in objective, tangible,
reachable concepts. It definitely oversimplified, so that even four decades ago,
basic research on the dimensions (factors) of the mind had expanded far beyond the
IQ, and the number of dimensions discovered is now way, way beyond the IQ. IQ
measures only small fish and has failed to catch and hold onto the truly big fish that
the word “intelligence” signifies to most people. The general public has been
misled and still readily believes that IQ is the “total intelligence” that encompasses
all of the dimensions (factors) of the mind, including all creative factors (and
possibly all other dimensions not yet discovered and measured in the total mind-
power).
This measured concept of intelligence is much narrower than Dr. Machado’s
recent concept of the “development of intelligence,” which was included in his
minister’s title in the Venezuelan cabinet. Along with the program he initiated in his
country, he energetically sought teammates for his concept in many nations around
the world. Far more than being concerned just with increasing IQ scores of students,
he was essentially talking about developing the total potential mindpower of each
and every person. He emphasized the real powers of the mind, especially the
creative and other thinking and producing powers, not merely the less powerful
processes of learning from the library of knowledge of the past.

Four major approaches to creativity

According to Ross Mooney (1963), there are “four significantly different ap¬
proaches to the problem of creativity, depending on which of four aspects of the
Approaches to and definitions of creativity 101

problem a person uses to gain his initial hold.” Mooney’s phrasing about a person
gaining his initial hold suggests that each person may have his own explicit or
implicit definition or subdefinition of creativity or even definitions of one or more
subcomponents of creativity going along with his approach in order to get an initial
start on the problem of creativity. Mooney’s four approaches are
1. the environment in which the creation comes about, that is, the creative environ¬
ment (or climate or situation or place),
2. the product of creating, that is, the creative product, or
3. the process of creating, that is, the creative process, or
4. the person who is creative, that is, the creative person.

In all the creativity research conferences, and also in the work of our own Utah
team, all of these four approaches have been found to some degree as an approach
or set of approaches used by investigators in their projects and/or programs of
research. As research findings have started to accumulate in each of these four
areas, some predictive equations have been set up and validated. The creative
process and the creative product have typically been seen as the criteria of
creativity; the creative person has been the main basis of the predictors in the
equation; and the environment has been used variously as a modifier in the equation
as well as the stimulus situation through which the inner creative processes are
activated.
Both the conferences and our work have been strong in the criterion area, es¬
pecially in multiple measures of the creativity of products, including performances,
behaviors, publications, ideas, and other types of accomplishments. We are possi¬
bly as well prepared as anyone to write a full article on the complexity of the
criterion problem. In one conference, possibly the only one on the general problem
of the criterion, we definitely had the most data and the widest variety of experi¬
ences in working on criterion problems (Henry, 1967). In our own projects we have
worked on the creative processes as belonging in the criterion area in some cases
and in the person-predictor area in other cases, such as in our development of a
Creative Process Check List.
In covering these four approaches in greater detail, we are focusing not only on
basic research studies but also on implementation efforts. These include research on
implementation in schoolrooms and in other settings in education, business, indus¬
try, and government organizations.

The creative environment

Typically this includes the total complex situation in which the creative processes
are initially stimulated and sometimes sustained through to completion. The en¬
vironmental situation can be a natural or a typical environment. It can even be one
in which very deliberate attempts have been made to design a total environment,
such as specially designed instructional media used to spark and sustain creative
processes in one or more individuals.
102 C. W. Taylor

Our first in-the-world Architectural Psychology Program could be used to design

total collective environments for this purpose. One of the program’s slogans is
“Peace of Mind from the Well-Designed.” This could alternatively be stated as
“Activating a Creative Mind from the Well-Designed.” For example, the large
Salk Institute building in La Jolla was described by a San Diego newspaper as “An
Architectural Experiment in Creativity.”
In our Sixth Creativity Research Conference report, Instructional Media and
Creativity (Taylor & Williams, 1966), the first two chapters were entitled “creativ¬
ity through Instructional Media: A Universe of Challenges (Parts I and II).” This
report was drawn from notes about the impact of instructional materials on the
creativity of students. A total of 36 of the 42 pages in my first chapter contained
discussions (with each person named when sparked to comment from the challenges
in the think piece). It provided a marvelous kick-off session for the conference,
because all 14 persons around the table entered into the discussion multiple times. A
shortened and further edited example of this discussion from pages 16-19 in In¬
structional Media and Creativity (Taylor & Williams, 1966) follows, with the
discussants being identified by name:

Rogers: The “mind set” of the person doing the teaching is so important. A leading physicist
at the California Institute of Technology has been very much involved in trying to free up
their teaching of physics. In addition to the rather standard material, he purposely put in his
new text lots of little peripheral issues - puzzles that no one knows the answer to - to get the
students thinking about them. Now there is a strong tendency for the graduate assistants who
teach the quiz sections to include quiz questions on these peripheral things just to stimulate
students and let them feel that there are a lot of unanswered questions.
Taylor: There are both knowns and unknowns in a field. We must learn somehow to give
attention in the classroom to the unknowns as well as to the knowns, even though this has
hardly ever been done to date.
Hughes: I am convinced that teachers will have to acquire new concepts of their jobs and new
concepts of their relationship to students in order to eventually do any of these new things.
One of the things research shows is that, in general, teachers behave very much the same in
all American classrooms. Current concepts of what a teacher does are much more similar than
dissimilar. The personality factors of teachers that have been investigated do not show a
relationship to teaching pattern - they just show that teachers work under a cultural, soci-
etally defined set of expectations. In some way the basic concepts of the role of the teacher
will have to be reevaluated or restated.
Beck: Does this suggest that we train teachers according to a pat formula, and then we put
them into cells where they operate for the rest of their lives with practically no feedback?
Hughes: Yes, there is very little feedback as long as they keep control of the classroom - as
long as the classroom doesn’t go to pot.
Beck: Then they have no opportunity to watch others teach. Each one is in his or her little cell
working alone and working according to the ways that he or she has been taught.
Hughes: I would even go further than that. Our training of student teachers has practically no
effect later on what they do later as teachers in the classroom. How they teach in the
classroom emerges from the concept they have gotten as a teacher, through the years in
school under teachers plus the cultural stereotypes and cliches and so on, rather than through
the training that as student teachers they obtained in colleges and universities.
MacKinnon: Are you suggesting that as they later teach, they do something quite different
from what they have been taught to do as student teachers in training?
Approaches to and definitions of creativity 103

Hughes: Yes. Indeed they do something quite different from what they have been taught
about teaching. But instructors in education, in educational psychology and so on, teach as
they have been taught, too, but they tell their student teachers to teach in a different way than
they themselves teach.
Guilford: The student teachers use their own teachers as models of how to teach.
Hughes: Yes, and they are responding to the models.
Rogers: At Wisconsin we tried an experimental program where we really taught the student
teachers quite differently, where they gained quite a different kind of experience than occurs
in typical teaching. The data still are very tentative on this thing. Many of them found that the
kind of teacher they had been taught to be couldn’t live in the educational subculture; so they
dropped the way they were taught to teach and simply became conventional teachers because
there was no place for them.
Taylor: That’s the same as for supervisory development programs. You deliberately take
supervisors out of their work situation and train them. But when you put them back into the
old situation, the old system generally overpowers them and their training.
Edling: What I think is unique about new media is that we are introducing something new
into the system, and it makes this new kind of a role for a teacher more possible, provided
that the new media come in sufficient quantities for every teacher to use in the entire system.
Fames: A situation at our institute relates to this idea that new media might provide a new
opportunity for teachers who might not otherwise be willing to change. One man in particular
came through our 3-day program where we taught our methodologies of creative problem
solving, and then he participated in a follow-up 2-day program where we have them teach
novice students the same concepts. At the end of the first 3 days he didn’t think that the whole
business was worth a dam. He didn’t see anything in it. But he said that when he taught this
to students and when he observed what happened to them, it opened up his eyes completely to
this approach. I think the new media might tend to help the teacher do this.
Williams: We have been experiencing in film learning for a number of years. It is not the
passiveness of the observer when he observes someone else teach by new approaches, but the
activeness of the person himself when he takes over to do this new type of teaching.
Parnes: Yes, but it wasn’t so much this man’s active participation in teaching as it was his
observation of what was happening to the students when they were learning it. Initially he
didn’t think there was anything to the methods, but the final 2 days he reevaluated his
opinions when he saw that something different was happening to the students.
Williams: In designing new media devices, then, the point may be how to get the classroom
student and even the teacher himself active in observing what happens to students rather than
passively showing a film.
Hughes: What experience did this man have that made him sensitive or allowed him to be
sensitive to the reaction of students? Because this is one of the things to which our teachers
are generally not sensitive. And so I am interested in how he became aware of the conse¬
quences of his teaching in terms of what was happening in the students.
Torrance: That gives me another idea about our experience. Maybe what we have put into
our own experiment of having the teachers write their own reactions and observations of the
pupils’ reactions might have been more powerful than anything else we did, including having
the teachers write their own reactions to their new teaching experiences.
Hughes: This is a sensitizing process that is sorely needed.

Three of the 10 articles in Clues to Creative Teaching (Taylor, 1963-1964)

contribute to the environmental approach. One article points out that existing library
knowledge may be an asset or a deterrent to creativity. Furthermore, accumulation
of past knowledge by a person certainly provides no assurance that such a person
will be a producer of new knowledge beyond that which is known.
104 C. W. Taylor

These discussions and my research show that past knowledge can be learned and
retained with such great force that a person might be unlikely to stretch out beyond
that knowledge into the unknown or even to break away from that knowledge to
revise it or to produce future knowledge that has not yet been produced. Converse¬
ly, other persons may use past knowledge as stepping-stones - standing on the
shoulders of past knowledge. One can thereby rise above the past level to create
new knowledge at a higher level. A step toward creating an atmosphere conducive
to new knowledge could be to have a unit of existing knowledge be presented in a
provocative way so that students would process that knowledge creatively. The
stimulus nature of the input knowledge and the method used in its presentation can
be important in whether such knowledge is an asset or a liability in activating
creative processes in a person.
Along with this issue of the nature of stimulus and form of presentation is the
issue whether the person’s mode and channel of reception are favorable or unfavora¬
ble toward creative use of environmental stimuli. The first input stage can help
determine whether or not the inner processes will start to be creative. These are
dealt with, to some degree, in two of the Clues to Creative Thinking articles,
“Learning and Reading Creatively” and “Listening Creatively” (Taylor, 1963-
1964).
In these two articles we talk about marginal reading, which can produce marginal
thinking - a play on the ambiguous word “marginal.” In reading a printed page
creatively, a person can write notes in the margin of the printed page either about
the message on the printed page or about ideas of one’s own that are sparked from
reading the page. In the latter case, the person considers the printed page to be
“what has [already] been written,” and the marginal notes can later be expanded
into another column of “what is now being written by oneself that has not been
written before.”

The creative product

The products of creativity can include behaviors, performances, ideas, things, and
all other kinds of outputs, with any of all channels and types of expressions. The
criteria of creativity define the targets against which a predictor or batteries of
predictors are validated through correlation and multiple regression methods.
Our first major criterion study entailed 52 different measures of career contribu¬
tions of U.S. Air Force scientists, including eight different sources from which data
on each scientist were collected. These included 10 ratings and 1 ranking by imme¬
diate supervisors; 4 ratings by laboratory chiefs; 6 ratings, rankings, and nomina¬
tions by peers; 12 quantity (from vitae) and quality (judged by senior scientists)
scores on publications and research reports from the researchers who worked on the
project; 1 score on membership in professional societies; and 5 “control” variables
(to determine if they influenced any of the preceding criterion scores).
These 52 scores were factor-analyzed to yield 14 criterion factors (plus one
Approaches to and definitions of creativity 105

control factor in a 15th separate dimension). The rotated factor analysis matrix is
presented in Scientific Creativity (Taylor, Smith, & Ghiselin, 1963, pp. 60-61).
In further analyses of these U.S. Air Force project data in Widening Horizons in
Creativity and in the Seventh Creativity Research Conference report on Climate for
Creativity by Taylor and Ellison (1964, 1972), 130 predictor scores derived from 14
different tests were validated against each of 14 criteria. The predictabilities for
these 14 criterion factors mentioned earlier are listed in rank order below. The
percentage of the battery test scores found to be valid against each criterion, in turn,
is listed in parentheses. In addition to the results below, the predictabilities of three
of the initial (unfactored) criterion scores were 35% for Peer Rankings on Produc¬
tivity, 29% for Supervisor Ratings on Creativity, and 25% for Supervisor Ratings of
Drive-Resourcefulness.
Likableness as a research team member (44%)
Scientific and professional society membership (43%)
Current organizational status (38%)
Judged work output (35%)
Supervisory ratings of overall performance (35%)
Productivity in written work (32%)
Creativity rating by laboratory chiefs (29%)
Originality of written work (20%)
Visibility (20%)
Recognition for organizational contributions (17%)
Recent publications (14%)
Contract monitoring load (11%)
Status-seeking, “organization-man” tendencies (8%)
Quality (without originality) of research reports (2%)

Six of the foregoing criterion factors were selected as separate targets using
multiple regression equations. Appropriate indices were also computed to analyze
the percentage of criterion variance overlapped by each of the 52 scores in the
predictor test battery, as reported in a later section. A somewhat different set of six
criterion factors was also used to derive subscore predictors and total score predic¬
tors from a specially constructed lengthy biographical inventory, as also described
later.

Process versus product in creativity: a spontaneous discussion

Consider the following edited discussion of process versus product in creative

thinking (Taylor, 1964):

Barron: It has been assumed in most of our discussion that we can determine whether or not a
person is creative by observing his or her behavior or discovering what his or her products
are. This kind of definition is probably basic to the kinds of prejudices that psychologists
have. One could just as well construe creativity as an internal process continually in action
but not always observable - or perhaps in some cases fundamentally unobservable.
Fiedler: Yes, but creativity surely must be identified eventually by its product - no?
Barron: No.
106 C. W. Taylor

Hyman: You mean that a person can go through life being creative and nobody will be able to
identify him or her?
Barron: Yes, by this type of definition.
Fiedler: But not by mine.
Hyman: But how are we going to be able to try to identify creativity, then?
McPherson: There are at least 28 definitions, expanding almost toward or beyond 50 or
more. Take you choice. [See the appendix to this chapter.]
Westcott: What do you mean by an internal process, as opposed to products or the results of
processes?
Barron: I think of it as something that is happening in the central nervous system. My own
basic interest in research on creativity stems from the hope it offers that one may find in
psychic creation the same formal variables that can be used to describe creative process in all
of nature.
Hyman: There are several different gradations between “process” and “product.” A lot of
good work has been done toward measuring cognitive structure operationally. The product,
then, could emerge from different cognitive structures. You could see this if you had effec¬
tive ways to index changes in the meaning and structure of the individual’s experience. I
think these changes can be indexed with present techniques. In this way we may be getting
very close, actually, to what Barron is talking about.
Westcott: One place where this might be very “studyable” would be in the development of
children, who, as they grow up, have to discover for themselves and invent for themselves
the millions of things that all adults know anyhow. It is my feeling that children are doing this
constantly. The processes whereby they reach the new views of the world are very much, I
would think, like the same processes whereby a scientist reaches a new view of his field.
Mullins: You say that no product is necessary for creativity. How do you define product? Is it
some useful product, or is there no sign of a product?
Barron: It would be unusual to find no evidence of creativity in behavior even though
creative process was occurring, but I would argue that this sometimes happens - that is, that
no sign of it appears.
McPherson: Even to the person himself or herself?
Barron: He or she might not be able to compare himself or herself sufficiently with others to
see the signs.
McPherson: Do you think that he or she might be creative but not know that he or she is
creative?
Barron: Yes.
Fiedler: But how do you operationally define this?

Barron: I cannot, without considerable further thought, offer an operational definition of

creative process occurring in the central nervous system without evidence in behavior and
without the fact of its occurrence being known subjectively. Yet from analogy to other
internal processes of which we are unaware, I shouldn’t think this to be impossible in
principle.
Parnes: But wouldn’t the person know that something is happening, whether he or she would
call it creativity or not?
Leary: I can’t understand how there can be any question about what Barron is saying. I can’t
understand how anybody would disagree with him.
Guilford: It is not very clear what he means.

Levine: Sprecher reports that somebody asked an engineer about his creativity, but he didn’t
consider himself as being creative. Everybody else said that he was. His approach was novel.
He was solving problems. But the engineer didn’t think that he was creative - he was just
doing his job.
Approaches to and definitions of creativity 107

Sprecher: He said, “I’m doing things which are routine to a scientist or a knowledgeable
person with my background. These other people think I’m wonderful in producing novel
ideas.” He had a good opinion of himself, but he said, “The things that they praise about me
are developed.”
Westcott: I think that the only thing we can say is that if a person does put out productive
creative materials, then it seems possible that he’s indulging in creativity. But if he doesn’t do
this, or you don’t see any of these products, if he doesn’t manifest any of this outwardly, then
you really don’t know what to say about him.
Barron: Yes, there are virtually no false positives in identifying “creatives,” but there may
be a great many false negatives (true creatives not identified).
Sprecher: People can produce creative acts or novel acts without having creative thinking. I
would expand my definition of creativity (I’ve seen lots of shaking of heads around here, but
you’re all wrong) to say that a person can be creative in the sense that he produces a novel
product by a very routine, plodding, technical way, with perhaps little thought.
Mednick: What I mean to say is that recognizing the usefulness of a product is an important
part of the creative act. The question that no one asked me when I discussed my notions
relating to creativity was, “How is it that a person who is creative can pick the appropriate
combination of associations from the many, many he has presented to him?” It seems to me
that this is a crucial question.
Hyman: If you read Edison’s story, you’ll see that it’s not trial and error.
Sprecher: Yes, it is. It’s a very deliberate research of many substances. Well, it’s not trial and
error in that sense, but it’s a technician’s search.
Hyman: Edison’s criterion was a complete system of what he needed to find. He had a very
clear-cut, planned strategy.
Sprecher: No. I’ve read Josephson’s biography of Edison, too, if that’s what you’re talking
about. I’m not referring to the fact that he had a grand conceptual scheme which would
encompass more than the light bulb. I’m talking about the fact that in one specific sense he
had a problem to solve: “I want this dam filament to last.”
Hyman: He had all the specifications of exactly what he wanted in that filament, too.
Sprecher: Then he produced this novel, valuable product in a routine way by just collecting
all possible materials in encyclopedic fashion.
Hyman: But first he formulated the complete problem of exactly what he wanted and how he
would recognize the product even before he saw it. It is this formulation of a context for
searching and for evaluation of his trials that I want to emphasize.
Leary: I venture to predict that creative performance or the production of creative expressions
will increasingly be seen as a systematic combining of elements which have not been
combined before. Many people who are seen as creative writers, artists, and so forth are
technicians who have wanted to be creative, wanted to play the creative game, and fiddled
around rearranging symbols. Henry Miller’s chapter in Ghiselin’s book (1955) is a very nice
account of how this author went at the technical problems of being a creative writer. He
experimented extensively with imitations of other authors. Finally, he quit imitating, plunged
deeply into the “ocean of reality” on his own, and went on to what I would call a creative
experience. [A few short exchanges about machines and creativity next occurred.]
Holland: We have a place in Chicago where you buy your own paints and drop them down on
a spinning wheel. You can get some very nice pictures.
Hyman: Leary, the point that you’re making is that if you can explain the process a man uses
and duplicate it with a machine in some way, then that’s not creativity. You feel that
creativity is some kind of - well, whatever it is I don’t know - but something that man can’t
simulate on a machine, is that it?
Leary: I’m distinguishing between the (inner) creative experience (or the inner creative
process) and the outcome of a so-called creative product.
108 C. W. Taylor

Harmon: Couldn’t you perhaps duplicate these same processes on a machine and, if you did a
good simulation, have a creative process - when there’s a human being involved?
Barron: That’s a critical point. If you can get a proper, full description of the creative
process, can it be entirely mechanical?
Westcott: Isn’t it true that a machine which produces some of these kinds of things uses
almost infinitely more work than a person does in accomplishing a comparable end? From
what we know about the machines, they can whir . . . and in fifteen seconds do five years’
work.
Barron: But only on what’s fed into them. We do more work than anything else in the
universe - we are the biggest workers.
Leary: We have experimented by using cut-up recombinations with convicts who’ve had
mystical experiences and who can’t report them verbally because they don’t have adequate
vocabulary. We can sometimes come out with very striking images which reduce this tremen¬
dous gap between experience and the expressive communication of it which has bothered all
of us for a long time.
Mednick: I think that the role of random behavior in the construction of this “cut-up” poetry
is a bit exaggerated. The creative act in cut-up poetry occurs once after the fiftieth or so
rearrangement of the cut-up words. The constructor has an “Aha!” experience and selects
this order of words as meeting some requirements. Incidentally, this is a way of assessing the
degree of creativity; that is, the more requirements a product meets, the more creative the
product is. Let me expand on the questions of requirements a bit more. I believe that we can
judge how creative a product is by the number of requirements it meets. This sounds very
simple-minded, perhaps, but it meets a minimum requirement itself in that it begins to be a
researchable definition. In terms of the definition, except under special circumstances, it is
difficult to specify what reliable requirements such products meet.
Sprecher: You are saying that this is creativity for Mednick. This is legitimate, so long as you
recognize it as yours, but allow other people to define creativity however they wish.
Beittel: Yes, that’s a good point. I was going to mention it. I can see why we perhaps study
architects and painters, because painters have set themselves a problem of constant innova¬
tion at present. I want to throw in one more thing. Our culture in the arts is heavy in this
expressionistic ego-centered feedback with the problem. I like the point that Murray made in
the essay in the Michigan symposium about what he called biotic creativity. He thought that
there was a way to transfer this into the social realm where you don’t have the ego-centered
relationship to the product. I think that type of transfer presents some hope as an image, for
the feedback would concern an almost pure process, and the product would be a little bit
sharable, in a sense.
Westcott: Certainly, in the realm of haikai poetry, a requirement of usefulness is not directly
met, so such poetry falls outside of this requirement of usefulness. I wonder whether this
criterion of usefulness does not fit in this case in the sense of the particular task that the artist
or the poet is trying to accomplish.
Mednick: If you can reliably specify this task, then you have a researchable problem. It seems
likely to me that often the aim of a particular artistic work was to communicate a given
emotional state to a given group of people. In this case we can ask, Does the painting do this?
Yes. It does? It does it beautifully? Then it is wonderful.
McPherson: What about time? It might do it at one time and not at another.
Westcott: Yes, this is true.
Taylor: I’m sure that we can’t solve many things here, but there is merit in getting these
issues out on the table, as all of you have just done. I suggest you note our studies wherein we
obtained at least crude measures, separately, for processes and for products and found the
relationships between these different measures of processes and products. (Ghiselin, Rompel,
& Taylor, 1964; Taylor & Ellison, 1964)
Approaches to and definitions of creativity 109

^ Reproductive^

The routine, well-socialized person who

experiences only in terms of what he
has been taught and who produces
only what has been
produced before.

0 \
/ £ / The person who \
•o \
f aj
> r experiences uniquely and \ 0 \
The innovating performer sensitively outside of game
who experiences only in terms / x ^ s\ concepts (either by choice or >
~o
o
Q.
of the available categories but ( 2 /C 4 J helplessly by inability) but who s
who has learned to manipulate 3 is unable to communicate CJ
o' 1 or uninterested in
these categories in novel -
xj-
combinations. / ^ v communicating these
\ ^
\ &
\ / The person who
experiences outside
'V the conventional
/ £7
/ X /
\ T A* /
manner.
outside the limits of ego and
labels, and who has learned to
' develop new modes of communications,
or who can manipulate familiar categories in'
novel combinations or who can let natural
modes develop under his nurture.

Creative crea^.

Figure 4.1. Categories for diagnosing creativity in terms of performance and experi¬
ence. The diagnostic grid is defined by two axes: experience and performance. Each
axis runs a continuum from creative to reproductive. (From Leary, 1964, with
permission.)

This is the end of the discussion. The next few paragraphs and graphs are from
Leary’s chapter (1964).
Because there is creative experience (i.e., creative awareness) and there is creative perfor¬
mance, a multidimensional Diagnosis of Creativity is possible. Awareness can be creative -
our experience can be direct and fresh; or it can be reproductive, i.e., within the in¬
terpretative framework of the already learned, in which case we see only what we have been
taught to see.
Performance can be creative - we can produce new combinations; or it can be reproduc¬
tive - a repeating of old combinations.
When we oppose these two dichotomous continua of experience and performance in
orthogonal axes, a diagnostic circle is obtained. Four “types” of creativity are defined by the
four quadrants of the diagnostic circle shown below in Figure 4.1.

Type 1. The reproductive blocked (no novel combinations, no direct experience), which I
estimate to comprise about 75% of our American population.
Type 2. The reproductive creator (no direct experience, but crafty skill in producing new
combinations of old symbols), comprising, let us say, the most visible, successful
12% of our population.
Type 3. The creative creator (new experience presented in novel performances), of which
we can hope for 1% in any Golden Age.
Type 4. The creative blocked (new direct experience expressed in conventional modes), a
somewhat cryptic 12%.
These four “types” of creativity are obvious by definition and seem to require no elabora¬
tion. What does deserve amplification is the social perception of these four types. A person in
110 C. W. Taylor

v Reproductive^
°ck,e4
IB .e$ co'

Unimaginative, incompetent hack.

1 A

Competent, responsible,
ny / reliable worker. / \ o \
<D
O / Reliable nihilist, Psychotic, \ uo 1
\
4) / Bold initiator religious crank.
> insensitive, Eo
unsuccessful who wins ^ eccentric <D
innovator game recognitions \ Solid, reliable 1 who uses 'B
whose shock value but whose fame [ 2 4 person with a conventional 5
EL changes to crumbles as \ "deeP streak” ' forms for u
L> fads of expressing
4 morbid curiosity
as fads of \ performance / mystical XF
CA \ performance \ change. / 3A / convictions. / £XJ /
\ change. & /
cu
The truly creative giant
recognized by his own
age and the ages to come.

3B
The mad creative genius, the undiscovered far-out
crackpot creator who is recognized by later
generations as a creative giant.

Creative cteat^.

Figure 4.2. Schematic diagrams of social labels used to describe types of creativity.
Inner circle illustrates positive social labels, and outer circle negative labels. (From
Leary, 1964, with permission.)

any of these four quadrants can be seen as effective or as incompetent by his culture - and,
for that matter, by cultural subgroups.
If we divide each type into those labeled by their contemporaries as (a) effective and (b)
ineffective, we obtain the eight categories presented in Figure 4.2. These two-dimensional
circular grids can be used to plot test scores or content-analysis indices along the two
coordinates in order to diagnose the individual. Thus we can work not just in terms of eight
types but in terms of two continua which define a wide expanse of diagnostic space. The
same mathematical and psychometric methods apply. We can plot the location of scores on
check lists of creative experience and creative behavior executed by the subject or by
appraisers of this performance. Multilevel discrepancy indices can thus be calculated which
provide opportunities for differential approaches in the development of creative behavior.

The creative process

In all our approaches we have always been interested in what a supervisor or teacher
and other medium can do for workers and students. Also, we try to discover what
the surrounding environment can do and what kinds of persons it takes to have their
inner processes functioning creatively. My teaching style is to use provocative
methods and rewards, so that students are treated as thinkers in class, after which
they keep thinking, intermittently, between classes about the subject matter and
about the stretching of their own minds and mind powers. One student volunteered
that in my class he learned “not only what education is doing for us but also what
education is doing to us." This concurs with what Sam Proctor (1978) has empha-
Approaches to and definitions of creativity 111

sized: A mind is a terrible thing to waste. A high percentage of students’ minds are
not being developed and utilized in typical schoolwork; to challenge education to
improve, students could write a book titled “Don’t dwarf our minds!” It is also true
that a large part of anyone’s mind is a terrible thing to waste.
Brewster Ghiselin worked with R. Rompel (a graduate student) and me on con¬
structing and developing a Creative Process Check List with 225 adjectives as items
(Ghiselin, Rompel, & Taylor, 1964). There were different types of responses to
these items. The first was on “States of Attention” before, during, and after the
moment of a person’s shaping a new insight. The second measured “States of
Feeling” before, during, and after the moment of insight. We divided our scientists
by their criterion measures into those high in creativity but not in material recogni¬
tion and success and those high in material recognition and success but not in
creativity. The Creative Process Check List appeared to be a promising research
instrument for investigation of the psychological processes of scientists. It also
appeared likely to prove useful as a predictor of scientific accomplishment. The
check list could be improved if we would now use phrases as well as adjectives in
expanding and revising the check list items.
One of our most successful studies in the classroom was accomplished in Project
Implode in the Bella Vista School in Utah’s Jordan School District. “Implode”
appropriately means an internal explosion - in this case, in students. Consequently,
it called for a lot of internal action in the brain processes in the students. The
teachers and even the principal became effective in activating these inner processes;
so the students produced tangible products, first in quantity, and then with some
traditional quality and occasionally even some creative quality features. Practically
all of the students became quite effective in at least one or more of the high-level
brain-power talents being activated and developed.

The creative person

Intellectual, personality, motivation, and group indicators provided the predictor

scores against the previously cited criterion measures in our validation equations.
Generally, creative thinking has been developed much more than have creative
characteristics. The latter is an area as yet remarkably untried in schooling.
In this section it is appropriate to talk about research by Taylor and Ellison
(1964), in which we found that the product of beta weights and validity coefficients
for each of 52 predictor scores showed the percentage of the variance of the criterion
target overlapping with each predictor score. The four most creative criterion targets
and two other criteria with some creative features were selected for further valida¬
tion work.
All of the 52 scores in the battery, in combination, overlapped approximately half
or more of what each criterion measured (i.e., the criterion variance). These per¬
centages were 50.0%, 46.8%, 70.7%, 71.0%, 66.0%, and 65.4%. The overlapping
of each criterion was accumulated mainly from a large number of small contribu¬
tions from the 52 predictor scores. In fact, for a single score, the contribution of 6%
112 C. W. Taylor

is quite infrequent, and a contribution of 10% is definitely rare for any criterion.
Therefore, no single test score usually adds much to the total contribution. In other
words, the large total overlap is obtained through many small, separate, low contri¬
butions. Through these analyses, we have also learned to have considerable respect
for low validities, even in the range of .20, especially if we have enough of them. If
high validities are not being found and if low validities are all that are available,
then we would like to have as many of these low validities as we can get.
Our well-developed and continually improved “dynamic” biographical invento¬
ries yield scores that have proved to be the best predictors by far and to cover the
most relevant number of dimensions of intellectual and nonintellectual attributes in
our test batteries. Our factor analyses of items in biographical inventories discover
great complexity in the number of separate dimensions that such inventories mea¬
sure. In our Form U inventory (Taylor & Ellison, 1983) the creativity score mea¬
sures from 7 to 10 different subcomponents or subdimensions of the total creativity
score. The research on biographical inventories suggests that it takes what we call
“a lot of little oars, each pulling its share in the general direction of the target, to
build up high validity coefficients.” Collectively these little oars bring about a
strong total pull in the right direction. However, no one of these oars usually
produces much validity by itself; so we advise persons not to lean too much on any
single oar.
The most sound approach and the one recommended here is to use first and
foremost the predictive efficiency of a complex, well-developed, valid, biographi¬
cal inventory. Then, afterward see how much other types of tests (or measures) of
intellectual talents, special abilities, grades, achievement test scores, and person¬
ality scores may add to the predictive efficiency of the biographical inventory
scores. Any early measurement of creative potentials should also involve a complex
battery of scores in order to account for a very high percentage of all that is
involved. In other words, one may need to measure 10, 15, 20, or more dimensions
in any predictor battery if one wants to account for very much of any performances,
products, and processes that are highly creative.
These analyses argue that creative performances are very complex and that no
single-variable test, no single-variable training program, no single-variable change
in working conditions, and no single-variable theory of creativity will account for
much of the total phenomenon unless the single variable happens to be a mighty
complex one in itself. In other words, our evidence from multiple studies argues
strongly that creative performance is a very complex multivariable phenomenon.
In addition to the definitions in this chapter’s Appendix (Repucci, 1960), we
want to report our own uses and experiences with definitions of creativity, with help
from Arnold Toynbee of Great Britain, Brewster Ghiselin of our university, and
Robert Lacklen of NASA. Fortunately we received permission by correspondence
to publish in Widening Horizons in Creativity (Taylor, 1964) a challenging question
by Arnold Toynbee (1964): Is America neglecting her creative minority?
Toynbee declares that
Approaches to and definitions of creativity 113

to give a fair chance to potential creativity is a matter of life and death for any society. This is
all-important, because the outstanding creative ability of a fairly small percentage of the
population is mankind’s ultimate capital asset. . . .
Creation is a disturbing force in society because it is a constructive one. It upsets the old
order in the acts of building a new one. This activity is salutary for society. It is, indeed,
essential for the maintenance of society’s health; for the one thing that is certain about human
affairs is that they are perpetually on the move, and the work of creative spirits is what gives
society a chance of directing its inevitable movement along constructive instead of destruc¬
tive lines.
A creative spirit works like yeast dough. But this valuable social service is condemned as
high treason in a society where the powers-that-be have set themselves to stop life’s tide from
flowing. This enterprise is foredoomed to failure. . . .
In present-day America [or so it looks to Toynbee] the affluent majority is striving desper¬
ately to arrest the irresistible tide of change. It is attempting this impossible task because it is
bent on conserving the social and economic system under which this comfortable affluence
has been acquired. With this unattainable aim in view, American public opinion today is
putting an enormously high premium on social conformity; and this attempt to standardize
people’s behaviour in adult life is as discouraging to creative ability and initiative as the
educational policy of egalitarianism in childhood.
Egalitarianism and conservatism work together against creativity, and, in combination,
they mount up to a formidable repressive force. . . . America rose to greatness as a revolu¬
tionary community, following the lead of creative leaders who welcomed and initiated timely
and constructive changes, instead of wincing at the prospect of them. ... It is ironic and
tragic that, in an age in which the whole world has come to be inspired by the original and
authentic spirit of Americanism, America herself should have turned her back on this, and
should have become the arch-conservative power in the world after having made history as
the arch-revolutionary one.
What America surely needs now is a return to those original ideals that have been the
sources of her greatness. . . . America’s need, and the world’s need, today, is a new burst of
American [creative] pioneering, and this time not just within the confines of a single conti¬
nent but all round the globe.
America’s [needed spirit and her] manifest destiny in the next chapter of her history is to
help the indigent majority of mankind to struggle upward towards a better life than it has ever
dreamed of in the past. ... If this spirit is to prevail, America must treasure and foster all the
creative abilities that she has in her.

Later, after Toynbee spoke at my 1967 summer creativity workshop on our

campus, he wrote us on April 18, 1968, a letter from London that

my impression - and indeed my conviction - is that, if America is to treasure and foster all
the creative ability that she has in her, a new and right spirit of change has to be injected into
her educational philosophy. The rather rigid egalitarian models of educational selection and
treatment will, I should press, have to be refashioned to include the creative talents of the
coming generations.
It would follow from this, if I am right in my diagnosis, that new educational philosophies
and new institutions of learning need to be constructed to provide an opportunity for creative
individuals to enhance their talents in schools. If the American people, or any other people,
are unwilling to change their minds and hearts to remold their educational establishment in
ways that foster creative talents, they cannot expect to be able to persist in this negative
attitude with impunity. This is surely true, particularly of the American people, conceived,
nurtured, and guided, as it has been, for almost two hundred years by its creative leadership.
“Where there is no vision, the people perish.”
114 C. W. Taylor

Our first major study on creativity of scientists involved U.S. Air Force scien¬
tists. One member of our team was Brewster Ghiselin, who suggested that the
measure of the creativeness of a product should be “the extent to which it restruc¬
tures our universe of understanding” (Taylor and Ellison, 1964, p. 228). His
formulation, which is applicable across disciplines, was also suitable for our mea¬
surement of the performance and contributions of Air Force scientists. That is, the
more creative the contribution, the more it restructures one’s universe of under¬
standing. Conversely, the less creative the contribution, the less it affects and
requires any restructuring of our total universe of understanding.
In our studies of NASA scientists (Taylor & Ellison, 1964, p. 229) we used the
breadth-of-applicability definition, which Robert Lacklen, NASA’s first personnel
director, produced and which was widely accepted by NASA. The degree to which
a NASA scientist’s contribution is creative is found by determining the extent of the
area of science that each contribution underlies. The more basic a contribution, the
wider its effects. In the NASA creativity scale, the highest degree is as follows:
The impact of his work has been quite exceptional. His creative solutions to complex
problems have broad generality and have even opened up important areas of investigation
with wide implications.

NASA’s lowest degree of creativity is described as follows:

His work has demonstrated very little creativity or originality. It usually has provided no
more than a rather simple solution to the immediate problem.

To deal with the inner processes of individuals majoring in liberal arts on cam¬
pus, a pertinent restatement of Ghiselin’s definition would be as follows: “The
creative process is the process of change, of development, of evolution, in the
organization of subjective life.” This implies an internal restructuring of a student’s
own universe of understandings and insights - even a reorganization of one’s inner
established order, especially if it is too entrenched. Students who are most adept at
their own individual creativity are the ones who most easily break out of their
established order and restructure it as needed.
At our Third Creativity Research Conference (Taylor, 1959), Golovin and Kuhn
engaged in an unrecorded debate about whether definitions are prerequisites to
progress or outcomes of progress. Both had doctorates in physics, Golovin as the
space expert in the Advanced Research Projects Agency of the Department of
Defense, and Kuhn was becoming eminent as a philosopher of science from his
writings on the structure of scientific revolutions. The relation of definition to
progress is an important one, with Golovin taking the first position (definition as a
prerequisite) and Kuhn the second (definition as an outcome).
Recently, Joe McPherson, another major creativity researcher, remarked that
“closure is not a way of life to recommend to anyone.” Nor is it particularly wise to
recommend to any growing movement, such as the creativity research and imple¬
mentation movement. Increasing openness is more in tune with creativity and
longevity.
Once I thought I heard a person say “hopenendedness,” a word I liked because it
 PRODUCTIVE DECISION- PLANNING HUMAN DISCERNING
ACADEMIC THINKING COMMUNICATING FORECASTING MAKING (DESIGNING) IMPLEMENTING RELATIONS OPPORTUNITIES
116 C. W. Taylor

was full of hope. However, I checked with him and found that he had said ‘ ‘openen-
dedness” instead, which also provides opportunities for hope. Controls and re¬
strictions that limit the meaning and scope of creativity are contrary to the spirit of
creativity. They are also contrary to the characteristics found in creative persons,
such as flexibility, discernment of opportunities, tolerance of ambiguity, and avoid¬
ance of premature closure, especially rigid closure. The expanding field of creativ¬
ity could be stifled by curtailing divergence and diversity, by maximizing con¬
vergence, and by crystallizing too solidly. It would be an injustice to the current
lively movement and could result in failure to reach the highest level of creativity
that is now being sought after and being approached gradually and asymptotically.
At this time, any person interested in creativity can find a subcomponent or com¬
bination of subcomponents of creativity to get hold of and move ahead to make
whatever contribution possible.
Another relevant area of our work has been in teaching for multiple creative
talents beyond the academic talent, which essentially covers traditional teaching and
testing for knowledge. Initially, while still at the theoretical stage, we had two
talents in mind, namely, academic and creative talents, with the latter being the
breakthrough and breakaway talent. As we moved into classroom practices, we
retained the academic talent and replaced the second talent with the five talents of
productive thinking, planning, communicating, forecasting, and decision making,
which were shown in a set of theoretical totem poles as first sketched by Darrel
Allington.
In 1969, Beverly Lloyd first put 28 live second grade students, four sets of seven
at a time, on the six totem poles listed earlier. Later, she followed them through
high school as her dissertation research at our university (Lloyd, 1983). Figure 4.3
shows nine talent totem-pole illustrations, the first six described earlier plus three
new ones added in the last 4 years, namely, implementing, human relations, and
discerning opportunities. The last eight talents are all thinking, producing, and
creating types of talents.
After these analyses, we recently decided to do a synthesis, because all these
eight totem poles can be used to function creatively. A group of experienced
creativity researchers assigned weights to yield a Creativity Composite Score to
return to a two-totem-pole illustration of Academic versus Creativity, as seen in
Figure 4.4. The point here is that this is another way to use the last eight subcompo¬
nents (totem-pole scores) in combination to define and yield a Creativity Composite
Score for each student on the total set of talent totem poles. This classroom imple¬
mentation and research evaluation work has now been going on for nearly 25 years
in some schools all around the nation.
In summary, in our work on the creative environment (or place) we used from 7
to over 15 different dimensions or subcomponents in studying the “total” environ¬
ment in classrooms and large organizations. In our studies of the creative product
we covered up to 14 factored dimensions from a vast number of criterion scores. In
our creative process study we used over 200 items in two different ways to probe
initially into this complex, vital area. Finally, in our numerous studies of the
Approaches to and definitions of creativity 117

CREATIVITY
ACADEMIC COMPOSITE

ANN

DIANE

STEVE

RANDY

KATHY

TODD

LINDA

Figure 4.4. Totem poles for two broad talents. (Copyright 1984, Calvin W. Taylor.)

creative person we used seven or more dimensions in our biographical inventory

scores and tested the validity of up to 52 predictor scores in our large study of
predictors and criteria of U.S. Air Force scientists.
Each different approach involved somewhat different problems of what descrip¬
tion and measures to use. Great complexity was found in terms of the numbers of
dimensions in these different types of studies of creativity. Consequently, it pro¬
vides a great unmet challenge for anyone who seeks to produce one definition that
fits well and satisfies all these different approaches.
In conclusion, we hope to encourage the diversity of approaches necessary to
keep the creativity movement growing and thriving. We believe that the importance
and complexity of the highest level of creativity are worthy of a continually diversi¬
fying movement with increasing support. The movement should not miss the “big
creativity fish” and settle prematurely for a “little fish,” as the group-IQ move¬
ment seems in hindsight to have done.
We move toward Creative and not just Traditional Excellence, because the latter
focuses on perpetuating the past and defaults the future creative-history-making to
others elsewhere. We therefore insist on improvements through cultivating both
Creative Excellence and Traditional Excellence, not either one without the other.
118 C. W. Taylor

Appendix

Definitions of creativity

L. C. Repucci
Dow Chemical Company

A review of the literature on thinking and problem-solving reveals a variety of theoretical

orientations and a whole host of experimental investigations. To sift through this mass of data
is a separate task in and of itself. Consequently, we shall focus on one specific aspect of the
thinking-problem-solving dimension. This is the area referred to as creativity or creative
problem-solving.
In order to orient ourselves, we must briefly consider the semantics of the word itself. That
is, what is creativity? What does the word mean?
A great deal of effort has been spent to reach some understanding of the word. At present,
our investigation reveals the existence of some 50 or 60 definitions and list is expanding
every day. If we examine the many definitions which have been offered, we can classify them
into six major groups or classes. These groupings are not mutually exclusive since each
definition may contain elements which fall into different classes. To overcome this difficulty,
the class into which a definition was placed was determined by the main theme of the
definition.
The first class of definitions will be labeled “Gestalt” or “Perception” type definitions
wherein the major emphasis is upon the recombination of ideas or the restructuring of a
“Gestalt.” Certainly, Wertheimer’s (1945) definition that creativity is the “process of
destroying one gestalt in favor of a better one” belongs in this category. So also the definition
of Keep (1957) that it is “the intersection of two ideas for the first time” and Duhrssen’s
(1957) notion that it is the “translation of knowledge and ideas into a new form” belong in
this category. Mooney (1955), Von Fange (1954) and others utilize this approach.
The second class of definitions may be called “end product” or “innovation” oriented
definitions. A representative member of this class is Stein’s definition (1953) that “Creativity
is that process which results in a novel work that is accepted as tenable or useful or satisfying
by a group at some point in tie.” Even Webster’s dictionary (1953) is oriented in this
direction for “to create” is defined as “To bring into being,” “To produce as a work of
thought or imagination.” Harmon (1955) prefers to speak of it as “any process by which
something new is produced - an idea or an object, including a new form or arrangement of
old elements.”
A third class of definitions can be characterized as “Aesthetic” or “Expressive.” The
major emphasis here is upon self-expression. The basic idea seems to be that one has a need
to express himself in a manner which is unique to him. Any such expression is deemed to be
creative. Hence we have Lee’s definition (1957) that “the creative process can be defined as
ability to think in uncharted waters without influence from conventions set up by past
practices.” In this vein, Lange (1957) offers that “The creative process is God, the creator,
working through his creation, man.” Northrop (1952) speaks of it in terms of stimulating our
sensibility to dissatisfaction and Compton (1952) sees the essence of creativity as being the
decision to do something when you are irritated. Thurstone (1954) thinks of it in terms of
problem sensitization and Ghiselin (1955) defines it as “the process of change, of develop¬
ment, of evolution, in the organization of subjective life.”
A fourth class of definitions can be characterized as ‘ ‘psychoanalytic’ ’ or ‘ ‘dynamic. ’ ’ In
this group, we find creativity defined in terms of certain interactional strength ratios of the id,
ego and superego. In this respect, Beliak (1958) assumes that all forms of creativity are

First section of an unpublished 1960 report to Dow Chemical Company titled Definitions and
Criteria of Creativity.
Approaches to and definitions of creativity 119

permanent operant variables of personality and he subscribes to the notion that to be creative
the ego must regress in order for preconscious or unconscious material to emerge. Leading
proponents of this type of definition are Anderson (1959), Kris (1951) and Kubie (1958).
A fifth class of definitions can be grouped under the classification of “Solution Thinking. ”
Here the emphasis is upon the thinking process itself rather than upon the actual solution of
the problem. Spearman (1931), for instance, defines creativity in terms of correlates. That is,
creativity is present or occurs whenever the mind can see the relationship between two items
in such a way as to generate a third item. Guilford (1959) on the other hand, defines creativity
in terms of a very large number of intellectual factors. The most important of these factors are
the discovery factors and the divergent-thinking factors. The discovery factors are defined as
the “ability to develop information out of what is given by stimulation.” The divergent-
thinking factors relate to one’s ability to go off in different directions when faced with a
problem. This is similar to Dunker’s notion (1945) that in order to solve a problem one often
must move tangentially from common types of solutions. Other proponents of this class of
definitions are Poincare (1913) and Wallas (1926).
The sixth and last class of definitions is labeled “Vart'a” simply because there is no easy
way of characterizing them. There is, for instance, Rand’s (1952) definition that creativity is
the “addition to the existing stored knowledge of mankind.” Lowenfeld (1957) speaks of it
as the result of our subjective relationship with man and environment. Porsche (1955) sees it
as the integration of facts, impressions, or feelings into a new form. Read (1955) feels that it
is that quality of the mind which allows an individual to juggle scraps of knowledge until they
fall into new and more useful patterns and Shepard (1957) speaks of it as a destructive process
much like Wertheimer when he spoke of creativity in terms of destroying one Gestalt in favor
of another.

References

Anderson, H. H. (1959, May). External and internal criteria of creativity. Paper presented at a sym¬
posium on creativity, Midwestern Psychological Association, Chicago, IL.
Beliak, L. (1958). Creativity: Some random notes to a systematic consideration. Journal of Projective
Techniques, 4, 363-380.
Compton, A. H. (1952). Case histories: Creativity in science. In F. Olsen (Ed.), The nature of creative
thinking (pp. 23-31). New York: Industrial Research Institute, Inc.
Duhrssen, A. (1957). Discussant in The Arden House conference on motivating the creative process.
Harriman, NY.
Dunker, D. (1945). On problem solving. Psychological Monographs, 58, No. 270.
Ghiselin, B. (1955). The creative process. New York: Mentor.
Ghiselin, B., Rompel, R., & Taylor, C. W. (1964). A creative process check list: its development and
validation. In C. W. Taylor (Ed.), Widening horizons in creativity (pp. 19-33). New York: Wiley.
Guilford, J. P. (1959). Intellectual resources and their value as seen by creative scientists. In C. W.
Taylor (Ed.), The third (1959) University of Utah research conference on the identification of
creative scientific talent (pp. 128-149). Salt Lake City: University of Utah Press.
Harmon, L. R. (1956). Social and technological determiners of creativity. In C. W. Taylor (Ed.), The
1955 University of Utah research conference on the identification of creative scientific talent (pp.
42-52). Salt Lake City: University of Utah Press.
Henry, E. R. (Ed.). (1967). Conference on criteria for research and application of research in manage¬
ment, leadership and creativity. Greensboro, NC: Smith Richardson Foundation.
Keep, O. A. (1957). Discussant in The Arden House conference on motivating the creative process.
Harriman, NY.
Kris, E. (1951). On preconscious mental processes. In D. Rapport (Ed.), Organization and pathology of
thought (pp. 474-496). New York: Columbia University Press.
Kubie, L. S. (1958). Neurotic distortion of the creative process. Lawrence: University of Kansas Press.
120 C. W. Taylor

Lange, G. C. (1957). Discussant in The Arden House conference on motivating the creative process.
Harriman, NY.
Leary, T. (1964). The effects of test score feedback on creative performance and of drugs on creative
experience. In C. W. Taylor (Ed.), Widening horizons in creativity (pp. 94-96). New York: Wiley.
Lee, W. J. (1957). Discussant in The Arden House conference on motivating the creative process.
Harriman, NY.
Lloyd, B. (1983). Longitudinal effects of multiple talent training on 28 second grade students - the totem
pole kids. Unpublished dissertation, University of Utah, Salt Lake City.
Lowenfeld, V. (1957). Creative and mental growth. New York: Macmillan.
Mooney, R. L. (1955). Cultural blocks to creative activity. In J. E. Arnold (Ed.), Creative engineering
seminars. Cambridge: Massachusetts Institute of Technology.
Mooney, R. L. (1963). A conceptual model for integrating four approaches to the identification of
creative talent. In C. W. Taylor & F. Barron (Eds.), Scientific creativity: its recognition and
development (pp. 331-340). New York: Wiley.
Northrop, R. S. C. (1952). Philosophy’s statement of the problems of creativity. In F. Olsen (Ed.), The
nature of creative thinking (pp. 16-22). New York: Industrial Research Institute, Inc.
Poincare, H. (1913). Mathematical creation. In G. H. Halsted (Trans.), The foundations of science (pp.
383-394). New York: Science Press.
Porsche, J. D. (1955, October). Creative ability - its role in the search for new products. Paper
presented at special conference on managing product research and development, American Manage¬
ment Association, New York.
Proctor, S. D. (1978). A mind is a terrible thing to waste. Phi Delta Kappan, pp. 201-203.
Rand, H. J. (1952). Creativity - its social, economic and political significance. In F. Olsen (Ed.), The
nature of creative thinking (pp. 12-15). New York: Industrial Research Institute, Inc.
Read, G. M. (1955). Profile of human materials. Paper presented at a centennial symposium on modem
engineering, University of Pennsylvania, Philadelphia.
Shepard, H. A. (1957, August). The destructive side of creativity. The Chemist, pp. 303-307.
Spearman, C. (1931). Creative mind. New York: Appleton.
Stein, M. I. (1953). Creativity and culture. Journal of Psychology, 36, 311-322.
Taylor, C. W. (1959). The third (1959) research conference on the identification of creative scientific
talent. Salt Lake City: University of Utah Press.
Taylor, C. W. (1963-1964). Clues to creative teaching (featured 10-year series usually starting on p. 5).
The Instructor.
Taylor, C. W. (Ed.). (1964). Widening horizons in creativity. New York: Wiley.
Taylor, C. W. (Ed.). (1972). Climate for creativity. New York: Pergamon Press.
Taylor, C. W., & Barron, F. (Eds.). (1963). Scientific creativity: its recognition and development. New
York: Wiley.
Taylor, C. W., & Ellison, R. L. (1964). Predicting creative performances from multiple measures. In C.
W. Taylor (Ed.), Widening horizons in creativity (pp. 242—253). New York: Wiley.
Taylor, C. W., & Ellison, R. L. (1972). Predictors and criteria of creativity - a Utah progress report. In
C. W. Taylor (Ed.), Climate for creativity (pp. 149-165). New York: Pergamon Press.
Taylor, C. W,, & Ellison, R. L. (1983). Searching for student talent resources relevant to all USDE
types of giftedness. Gifted Child Quarterly, pp. 99-106.
Taylor, C. W., Smith, W. R., & Ghiselin, B. (1963). The creative and other contributions of one sample
of research scientists. In C. W. Taylor & F. Barron (Eds.), Scientific creativity: its recognition and
development (pp. 60-61). New York: Wiley.
Taylor, C. W., & Williams, F. E. (Eds.). (1966). Instructional media and creativity. New York: Wiley.
Thurstone, L. L. (1954, April). Criteria of scientific success and the selection of scientific talent. In W.
J. Brogden et al. (Eds.), Criteria of success in science (pp. 29-34) (Technical Report 4, Office of
Scientific Personnel). Washington, DC: National Academy of Sciences-National Research
Council.
Toynbee, A. (1964). Is America neglecting her creative minority? In C. W. Taylor (Ed.), Widening
horizons in creativity (pp. 3-9). New York: Wiley.
Approaches to and definitions of creativity 121

Von Fange, E. K. (1954). The creative process in engineering. General Electric Co., Creative Engineer¬
ing Program.
Wallas, G. (1926). The art of thought. New York: Harcourt, Brace.
Webster’s New Collegiate Dictionary. (1953). Springfield, MA: G. & C. Merriam Co.
Wertheimer, M. (1945). Productive thinking. New York: Harper.
■
Cognitive approaches
5 A three-facet model of creativity

Robert J. Sternberg

Alice should have been the ideal graduate student: She entered Yale’s graduate program with
800-level scores on the Graduate Record Examination, close to a 4.0 undergraduate academic
average, excellent letters of recommendation, and all of the apparent undergraduate socializa¬
tion that should have taught her what counts in academe. During her first year, when course
work, multiple-choice tests, and writing brief critical papers were stressed, Alice was every¬
thing her record promised she would be. But by the time she left the graduate program, her
record did not look promising at all. Alice continued to show excellent memory and critical-
thinking skills. But once having creative ideas started to matter, Alice proved not to be
competitive with even the typical students in the program. She was a good critical thinker,
but not a good creative thinker. Nothing in her undergraduate record prepared the faculty - or
Alice - for this unwelcome surprise. In contrast, Barbara, who entered the program with
relatively low test scores and only typical undergraduate grades, as admitted students go,
fared much better. Her critical ability was good, but it was not critical abilities that propelled
her to great success. Rather, she proved to be among the most creative researchers we had.
Only her letters of recommendation praised her creativity. There was nothing in the hard data
- test scores and grades - to indicate that Barbara would be an exceptional student.

In her book Whatever Happened to the Quiz Kids? Ruth Feldman (1982) describes the
childhood feats of the television quiz kids. These children dramatically impressed the world
with their extraordinary demonstrations of learning and recall. It was hard to believe that
young children could be so smart. Indeed, Feldman reported that all of them had very high
IQs. Yet, in reading their biographies, one cannot help but be struck by how much less
creative they were than they were intelligent. Certainly, their creative performances in their
later lives have nowhere near matched their intelligent performances in their earlier lives.

In the film Amadeus, both Mozart and his rival, Salieri, are portrayed as highly intelligent as
well as professionally ambitious. Had one given each of them a standard intelligence test
suitable for the times, it is unlikely that the test scores would much have distinguished their
relative levels of musical expertise. But Mozart was far more creative than Salieri, as Salieri
and Mozart both recognized. In the long run, Mozart was remembered, and Salieri all but
forgotten.

These three vignettes are only a small sample of the vignettes, as well as the data,
that indicate that for whatever overlap there may be between intelligence and
creativity, it is not complete. Reviews of the empirical literature (Perkins, 1981;
Vernon, 1970) reveal that creativity is unlikely to be understood merely as an

Preparation of this research was supported by Contract N00014-85-K-0589 from the Office of
Naval Research. Requests for reprints should be sent to Robert J. Sternberg, Department of
Psychology, Yale University, Box 11 A, Yale Station, New Haven, CT 06520.
125
126 R. J. Sternberg

adjunct of some other cognitive or personality phenomenon. Creativity overlaps

with other psychological phenomena, such as intelligence, cognitive style, and
personality, but it is not identical with any of them. The goal of this chapter is to
attempt to understand what creativity is and how it can be understood at a peculiar
intersection between three psychological attributes: intelligence, cognitive style,
and personality/motivation. Taken together, these three facets of the mind help us
understand what lies behind the creative individual.
The chapter will be divided into five basic parts. In the first, I review some
research that helps delineate the boundaries of creativity - what the phenomenon is
that needs to be understood. In the next three parts, I discuss those aspects of
intelligence, cognitive style, and personality/motivation that seem most to contrib¬
ute to the phenomenon that we label creativity. In the last part, I attempt to integrate
these various aspects of creativity into a tentative, unified account of this elusive
psychological attribute.

The nature of creativity

What is creativity? Few psychological constructs have proved more elusive to

define. Sometimes creativity is defined in terms of products, such as products that
are statistically unusual and also adjudged to be of high quality (Perkins, 1981). But
such a definition tells us little about the psychology of the creative person. In
studying phenomena that are not well understood, it is sometimes helpful to start
with a study of people’s implicit theories - or conceptions - of a given phe¬
nomenon. This is a course we took in the study of intelligence (Sternberg, Conway,
Ketron, & Bernstein, 1981) and have more recently applied to the study of intel¬
ligence, wisdom, and creativity and the interrelations among them (Sternberg,
1985b). Through such a study it is possible to understand what people perceive
creativity to be and how they see it as distinguished from intelligence and wisdom.
Such a study does not tell us what creativity “really is.” But it serves as a start
toward defining the core as well as the boundaries of the phenomenon.

Intercorrelational relations among creativity, intelligence, and wisdom

1 began (Sternberg, 1985b) by sending questionnaires to professors of art, business,

philosophy, and physics. The questionnaires asked these individuals to list behav¬
iors that were characteristic of people who were highly intelligent, wise, or creative
in their respective fields of endeavor. A comparable questionnaire was given to
laypersons - people from the New Haven area who answered an advertisement in a
local newspaper. The number of individuals responding averaged 25 per group. The
number of behaviors listed averaged 131 for each of the three attributes of intel¬
ligence, wisdom, and creativity, so that the listing procedure generated quite a large
number as well as variety of behaviors for each attribute.
Three-facet model of creativity 127

A new questionnaire was then sent out to different people within the same subject
population. These “new” people were asked to rate how characteristic each of the
set of behaviors was of an ideally intelligent, wise, or creative individual in their
respective groups. The rating scale ranged from extremely uncharacteristic (1) to
extremely characteristic (9). The average number of subjects per group was 65, and
all subjects provided all three ratings (for intelligence, wisdom, and creativity),
although in different orders. For example, some of the people would provide ratings
for creativity first, others would provide creativity ratings second, and still others
would provide creativity ratings last. The ratings obtained in each of the groups
proved to be highly reliable, meaning that there was substantial agreement among
the members of the various groups as to what constituted intelligent, wise, or
creative behavior.
The various ratings were intercorrelated in order to determine how closely related
the people in each of the groups thought intelligence, wisdom, and creativity to be.
Three findings of particular interest emerged from these intercorrelations.
First, I found that, in general, intelligence and wisdom are perceived as most
closely related (median r — .68), intelligence and creativity as next most closely
related (median r = .55), and wisdom and creativity as least related (median r
= .27). The only departure from this pattern was for philosophers, for whom
intelligence and creativity were considered to be more highly related (r — .56) than
were intelligence and wisdom (r = .42).
Second, all of the interrelationships were positive, meaning that greater amounts
of a given attribute were associated with greater amounts of each other attribute.
However, there was one exception to this trend: The business professors saw greater
amounts of wisdom as being associated with lesser amounts of creativity (r =
— .24). In other words, in business, the wiser people are seen as less creative, and
vice versa.
Third, there were some interesting differences in magnitudes of interrelations
across groups. For example, the art, philosophy, and physics professors all saw
intelligence and creativity as being highly related (r’s = .55, .56, .64, respec¬
tively), but the business professors and laypersons saw them as only weakly related
(r’s = .29 and .33, respectively). Moreover, the relation between creativity and
wisdom reached moderate levels in the opinion of the art professors and philosophy
professors (r’s = .48 and .37, respectively), but was low for the other groups and,
as mentioned earlier, was actually negative for the business group.
To conclude, the first study was critical in showing that whereas intelligence,
wisdom, and creativity are all viewed as being positively related, on the average,
the relations among them differ. Creativity is less closely related to the other two
constructs than these two constructs are to each other. Moreover, the positive
relation between wisdom and creativity actually becomes negative in the opinion of
the business group. Finally, the strength of the relation between intelligence and
creativity, although always positive, varied among groups. Business professors and
laypersons saw them as being less closely related than did the others.
128 R. J. Sternberg

The structure of implicit theories of creativity

In a second study, 40 college students were asked to sort three sets of 40 behaviors
into as many or as few piles as they wished on the basis of which behaviors are
“likely to be found together” in a person. These behaviors were from the listings
for intelligence, wisdom, and creativity from the first study. Only the top 40
behaviors (i.e., behaviors rated by laypersons as highly characteristic of ideally
intelligent, wise, or creative individuals) were used in each sorting task. Different
subjects did the sortings for intelligence, wisdom, and creativity in different orders.
A method of data analysis called nonmetric multidimensional scaling was used to
analyze the sortings. This method enables one to determine the latent structure in a
set of sorting or rating data. A latent structure was determined for each of the three
attributes. The scalings provided an excellent fit, statistically, to the data.
The multidimensional scaling for creativity yielded six major elements (Stress,
Formula 1 = .14. R2 = .87):
1. Lack of conventionality (e.g., one makes up rules as one goes along; has a free
spirit; is unorthodox).
2. Integration and intellectuality (e.g., makes connections and distinctions between
ideas and things; has the ability to recognize similarities and differences; is able to
put old information, theories, etc., together in a new way).
3. Aesthetic taste and imagination (e.g., has an appreciation of art, music, etc.; can
write, draw, compose music; has good taste).
4. Decisional skill and flexibility (e.g., follows gut feelings in making decisions after
weighing the pros and cons; has the ability to change directions and use another
procedure).
5. Perspicacity (e.g., questions societal norms, truisms, assumptions; is willing to
take a stand).
6. Drive for accomplishment and recognition (e.g., is motivated by goals; likes to be
complimented on work; is energetic).

For comparative purposes, it is useful to say something of the dimensions that

emerged for the constructs of intelligence and wisdom. For intelligence, six basic
elements also emerged: practical problem-solving ability, verbal ability, intellectual
balance and integration, goal orientation and attainment, contextual intelligence,
and fluent thought. For wisdom, six basic elements were found: reasoning ability,
sagacity, learning from ideas and environment, judgment, expeditious use of infor¬
mation, and perspicacity.
To summarize, there is some overlap, though it is far from total, in the perceived
dimensions of intelligence, wisdom, and creativity. The amounts of overlap reflect
the findings of the previously described correlational study: Intelligence and
wisdom show the most overlap, and creativity and wisdom the least. Most impor¬
tant, from the present perspective, the emerging view of creativity shows aspects of
intelligence, intellectual style, and personality/motivation. Lack of conventionality
will be viewed here as a matter of style - a way of approaching problems and
situations. The aspect of integration and intellectuality is defined as primarily an
aspect of intelligence, similar to the intellectual-balance-and-integration dimension
Three-facet model of creativity 129

that emerges for intelligence. The aspect of aesthetic taste and imagination contains
elements of intelligence, style, and personality/motivation. The aspect encompass¬
ing decisional skill and flexibility is a matter of intelligence and intellectual style, as
is perspicacity. Drive for accomplishment and recognition would seem to be pri¬
marily in the personality/motivational domain. Thus, it probably makes some sense
to attempt to understand the psychological antecedents of these aspects of perceived
creativity in terms of something like the three-facet approach used in this chapter.

Use of implicit theories of creativity in evaluating the creativity of

others

In the last of the series of studies, 40 individuals, all of whom were adults in the
New Haven area, were presented with 54 simulated letters of recommendation. Two
typical letters were as follows:
Gerald
He possesses ability for high achievement.
He has the ability to grasp complex situations.
He has good problem-solving ability.
He attaches importance to well-presented ideas.

Doris
She is motivated by goals.
She questions societal norms, truisms, and assumptions.
She thinks quickly.
She is not materialistic.
She is totally absorbed in study.

Descriptions were generated to vary predicted levels of intelligence, wisdom, and

creativity. Each description was either four, five, or six sentences in length and was
paired equally often with names of males and with names of females. A given
subject, however, saw a given description only once - either with a male name or
with a female name. The subject’s task was to rate the intelligence, wisdom, and
creativity of each of the described individuals. Ratings of the individuals were made
on a 9-point scale ranging from not at all intelligent, wise, or creative (1) to
extremely intelligent, wise, or creative (9). The order of ratings was varied across
subjects in the study.
It was possible to obtain predicted ratings of intelligence, wisdom, and creativity
by summing up the ratings of laypersons from the first study on each attribute for
each subject, and then dividing by the number of attributes given for the hypo¬
thetical individual. Averages rather than sums of ratings were used because the
number of behaviors was not the same for each of the descriptions.
Suppose, for example, that five behaviors were given for Susan. The predicted
intelligence ratings would be the average of the characteristicness ratings for intel¬
ligence in the first study (plus an additive constant). The predicted wisdom rating
would be the mean of the ratings for wisdom in the first study (plus an additive
constant), and the same held true for predicted creativity. Thus, the closer the
130 R. J. Sternberg

description of the hypothetical individual represents the ideal of the first study on
each of the three attributes of intelligence, wisdom, and creativity, the higher the
rating that the hypothetical individual receives in the present study.
Average ratings of the hypothetical individuals were highest for intelligence
(5.8), intermediate for wisdom (5.3), and lowest for creativity (5.0). The ratings
were highly reliable (median reliability = .98), meaning that there was agreement
among subjects as to who was relatively more or less intelligent, wise, or creative.
Intercorrelations (degree of relationship) of ratings were extremely high between
intelligence and wisdom (r = .94), high between intelligence and creativity (r
= .69), and moderately high between wisdom and creativity (r = .62). Thus, the
rank order of the correlational relations was the same as that in the past studies,
although in this study, intelligence and wisdom were almost indistinguishable. Use
of male versus female names had no effect either on level or on correlational pattern
of ratings.
The relations between observed and predicted ratings (with predictions deriving
from the first study) each showed an excellent fit of the predictions to the data. In
each case, the correlation between the predicted and observed values for a given
attribute was extremely high (median r = .89). Moreover, this correlation of pre¬
dicted with observed values for a given attribute (e.g., predicted values for
creativity with observed values for creativity) was always higher than the correla¬
tion of predicted with observed values across attributes (median r = .51) (e.g.,
predicted values for creativity with observed values for wisdom). Thus, people
seem not only to have implicit theories of intelligence, wisdom, and creativity but
also to use these implicit theories in predictable and somewhat distinguishable ways
to judge others.
To summarize, people have implicit theories, and they use these theories to
evaluate others. It is possible to predict people’s evaluations of others on the basis
of knowledge about their implicit theories. The results of their judgments closely
mirror the results of their mental structures as revealed by the previous correlational
and multidimensional scaling analyses.

Differences among laypersons and various experts in their conceptions

of creativity

To this point, implicit theories have been discussed primarily as though they are the
same from one group of people to another. Indeed, the studies described here show
a high degree of overlap between implicit theories across different groups. Howev¬
er, there are some interesting differences as well as commonalities. Consider these
differences for the analyses of creativity.
For laypersons, conceptions of creativity overlap with those of intelligence, but
there is much less emphasis in implicit theories of creativity on analytical abilities,
whether they be directed toward abstract problems or toward verbal materials. For
example, the very first dimension shows a greater emphasis on nonentrenchment, or
Three-facet model of creativity 131

the ability and willingness to go beyond ordinary limitations of self and environ¬
ment and to think and act in unconventional and even dreamlike ways. The creative
individual has a certain freedom of spirit and unwillingness to be bound by the
unwritten canons of society, characteristics not necessarily found in the highly
intelligent individual. Implicit theories of creativity encompass a dimension of
aesthetic taste and imagination that is absent in implicit theories of intelligence, and
they also encompass aspects of inquisitiveness and intuitiveness that do not seem to
enter into the implicit theories of intelligence. Implicit theories of creativity go far
beyond conventional psychometric creativity tests. A person’s abilities to think of
unusual uses for a brick, or to form a picture based on a geometric outline, scarcely
do justice to the kind of freedom of spirit and intellect captured in people’s implicit
theories of creativity.
For specialists, implicit theories of creativity in the specialized fields are highly
overlapping across fields and also highly overlapping with the implicit theories of
laypersons; nevertheless, there are some differences worthy of note. Professors of
art place heavy emphasis on imagination and originality, as well as on an abundance
of and willingness to try out new ideas. The creative artist is a risk-taker and persists
in following through on the consequences of risks. Such a person thinks meta¬
phorically and prefers forms of communication other than strictly verbal forms.
Business professors also emphasize the ability to come up with new ideas and to
explore these ideas, especially as they relate to novel business services and prod¬
ucts. The creative individual escapes traps of conventional thinking and can imagine
possible states that are quite different from what exists. Philosophy professors
emphasize the ability to toy imaginatively with notions and combinations of ideas
and to create classifications and systematizations of knowledge that differ from the
conventional constructs. Creative individuals never automatically accept the “ac¬
cepted,” and when they have novel hunches, these hunches often pay off. The
creative person is particularly adept at generating insights regarding connections
between seemingly unrelated issues and forming useful analogies and explanations.
Physics professors share many of these same ideas about the creative individual, but
show a particular concern with inventiveness, the ability to find order in chaos, and
the ability to question basic principles. The physicists emphasize creative aspects of
problem solving, such as the ability to approximate solutions, the ability to find
shortcuts in problem solving, and the ability to go beyond standard methods of
problem solving. Finally, the physicist thinks a creative person has the ability to
make discoveries by looking for reasons why things happen the way they do. Such
discoveries may result from the perception of physical and other patterns that most
others simply overlook.

Conclusions

What are the main conclusions to be drawn from this work on implicit theories?
First, people have implicit theories of creativity, as well as of intelligence and
132 R. J. Sternberg

wisdom, and they use these theories both in conceptualizing these attributes and in
evaluating themselves and others. Second, the three constructs are distinct, al¬
though interrelated. A complete theory of creativity would have to account for both
the overlap and the distinctness. Third, assessments of creativity are in need of
serious reconsideration and especially broadening. The measures of creativity pres¬
ently available to us measure creativity in a less meaningful (some might say more
trivial) way than would correspond to people’s implicit theories of creativity.
Thinking of unusual uses for a paper clip, for example, and similar tasks, would
seem to draw relatively little on the lack of conventionality, integration and intellec¬
tuality, aesthetic taste and imagination, decisional skill and flexibility, perspicacity,
and drive for accomplishment and recognition that are seen as essential aspects of
creativity. But then, how can we understand creativity so that we might better
measure it? The remainder of this chapter is devoted to developing a psychological
account of creativity that is consistent with, but in at least some respects goes
beyond, people’s implicit theories.

The intellectual facet of creativity

The intellectual facet of creativity refers to those aspects of creativity that can be
accounted for in terms of the theory of intelligence. The particular theory on which I
shall draw here is my own triarchic theory (Sternberg, 1985a). According to this
theory, intelligence can be understood as comprising three aspects: its relation to the
internal world of the individual, its relation to experience, and its relation to the
external world of the individual. It is through experience that the individual medi¬
ates the relation between the internal and external worlds.

The relation of intelligence to the internal world

In order to understand creativity, we need to understand the mental processes of the

intellect that underlie it. The triarchic theory of intelligence, through its componen-
tial subtheory, specifies some, although almost certainly not all, of these processes.
These processes can be executed in more or less creative ways. The creativity of the
way in which they are executed depends on the originality and quality of the
component execution. The relevant processes are of three kinds.

Metacomponents. Metacomponents are higher-order executive processes used in

planning, monitoring, and evaluating one’s problem solving. The metacomponents
most relevant to creativity per se are the planning metacomponents, or what I
sometimes call the “legislative” ones, in that they legislate what one is going to do.
1. Recognizing the existence of a problem. An important aspect of creativity is
the recognition of problems - not just any old problems, but problems that are (a)
large in scope, (b) important in their potential consequences, and (c) potentially
soluble, at least to some meaningful degree. Creative individuals are not just good
problem solvers but also good problem finders (Getzels & Csikszentmihalyi, 1976).
Three-facet model of creativity 133

They know a good problem when they see one, and indeed, they might best be
characterized as having “good taste’’ in problems (Zuckerman, 1983). In my own
experience of editing large numbers of manuscripts for a prestigious journal in the
field of child development, I found that what distinguished the more eminent from
the less eminent scientists was not always how well they designed or executed their
experiments but rather what experiments they deemed worthy of performing in the
first place. The greatest experiments are not always the most well-designed experi¬
ments. In psychology, for example, the seminal experiment on semantic memory,
that of Collins and Quillian (1969) on semantic networks, had a number of meth¬
odological holes and proved to be subject to a variety of alternative interpretations.
Similarly, the classic Shepard and Metzler (1971) experiment on mental rotation,
which was intended to show the analogue nature of mental representations, actually
yielded results that could be interpreted as consistent with a propositional represen¬
tation (Palmer, 1975); again, Tulving’s classic experiment (1966) on negative trans¬
fer in part-whole free recall, deemed to show the importance of organizational
effects in episodic memory, could be interpreted entirely in terms of list-discrimina¬
tion effects (Sternberg & Bower, 1974). What made these experiments great was
not the flawlessness of their designs but the vision that went into selecting important
problems, and then approaching them in what at the time seemed like a compelling
way. Problem selection was far more important than problem solution in determin¬
ing the classic role of these experiments in the history of psychology.
2. Problem definition. Once we have recognized an important problem, we still
have to define it. Sometimes we know we have problems facing us in life, but we
cannot quite figure out what the problems are, or what their structure looks like.
Problem definition requires the structuring of the problem in a way that makes the
problem both meaningful and soluble. In my own primary field of intelligence, for
example, what has changed over the past decade is only secondarily the way the
problems in the field have been approached. Rather, what has changed primarily is
the way the problems are structured. The goal of research is no longer seen as
understanding what conventional intelligence tests are measuring, but rather as
understanding what they should measure and how this differs from what they
actually do measure.
A classic problem showing the importance of problem definition - or redefinition
- can be found in the so-called nine-dot problem. In this problem, subjects are
presented with three rows of three dots and are asked to connect the dots with a
pencil, using no more than four straight lines and never lifting the pencil from the
paper. People unfamiliar with the problem usually define it in terms of connecting
the dots with lines that remain within the confines of the implicit border of the nine
dots. In fact, the problem can be solved only if one extends one’s pencil beyond the
nine dots. There was nothing in the statement of the problem to suggest that one
must stay within the implicit border of the nine dots. People tend to define the nine-
dot problem narrowly rather than broadly and hence miss out on the opportunity to
solve it at all.
Consider a second problem of a slightly different ilk:
134 R. J. Sternberg

A monk wishes to pursue study and contemplation in a retreat at the top of a mountain. The
monk starts climbing the mountain at 7:00 a.m. and arrives at the top of the mountain at 5:00
p.m. of the same day. During the course of his ascent, he travels at variable speeds and takes
a break for lunch. He spends the evening in study and contemplation. The next day, the monk
starts his descent at 7:00 a.m. again, along the same route. Normally, his descent would be
faster than his ascent, but because he is tired and afraid of tripping and hurting himself, he
descends the mountain slowly, not arriving at the bottom until 5:00 p.m. of the day after he
started his ascent. Must there be a point on the mountain that the monk passes at exactly the
same time of day on the two successive days of his ascent and descent? Why or why not?

The answer is affirmative, but it is extremely difficult to figure out if the problem
is defined in the way it is originally presented. The problem becomes much easier to
conceptualize if, rather than imagining the same monk climbing the mountain one
day and going down the mountain the next, the reader imagines two different
monks, one ascending and the other descending the mountain on the same day. In
this redefinition of the problem, the monk’s descent on the second day is reconcep¬
tualized isomorphically as a different monk’s descent on the same day as the first
monk’s ascent. If there were two monks, their paths of ascent and descent would
necessarily cross each other at the same time of day. Obviously, their meeting must
be at a given place and point in time, although one cannot specify in advance
exactly when or where the meeting will take place. Note that the problem becomes
easily soluble only when it is creatively redefined.
Creative redefinition of problems is not limited in its applicability or importance
to academic problems such as those cited. One instance relevant to many people’s
lives is what to do when one is involved in an intimate relationship and one is more
involved with the other person than the other person is with oneself. Most people
define the problem as one of equalizing levels of involvement by trying to get the
other person to be more involved - to reach one’s own level of involvement. Our
research suggests that such a strategy usually is nonoptimal (Sternberg & Barnes,
1985). A better strategy is to redefine the problem as one of more nearly equalizing
the levels of involvement by reducing one’s own level of involvement. Although
this reduction of one’s own level of involvement may be difficult to achieve, it is
more likely to lead to continuation and, ultimately, success of the relationship.
3. Formulation of a strategy and mental representation for problem solution. I
am not a great spatial visualizer. When I was younger, I would take tests of spatial
visualization ability and turn in some truly marginal performances. Somewhere
along the line, I realized that I did not have to perform at a marginal level on many
of these problems, because many of them could be solved verbally rather than
spatially. Once I had this realization, I no longer tried to solve the problems in the
way the test constructors intended. Rather, I would first try to determine if there was
a nonspatial means of solution and use this means of solution if I was able to find
one; I would use spatial processing only if no verbal strategy seemed to be avail¬
able. My scores on these tests improved substantially.
Many problems can be solved in multiple ways. Often, some of these ways are
more creative than others.
Three-facet model of creativity 135

Performance components. The performance components of intelligence execute the

instructions of the metacomponents. In most cases, the execution of these compo¬
nents is straightforward. However, in some instances, performance components can
be executed more or less creatively. For example, inferences can be more or less
creative, depending on how far beyond a given set of data one is able to go. A
creative inference is one that is unusual and high in quality or veridicality. Similar¬
ly, mappings - recognition of second-order relations (relations between relations) -
can be more or less creative, depending on their nontriviality.
Tourangeau and Sternberg (1981) have proposed a theory of metaphorical think¬
ing that is relevant to determining the level of creativity of an analogy. One can
imagine concepts as occurring in sets of local subspaces, with each subspace con¬
taining a set of related concepts (i.e., a semantic field). Thus, one can conceive of
the mind as comprising a large set of such local subspaces, which may overlap with
respect to certain concepts. In this conceptualization, it is possible not only to
compute the distance between points within a given local subspace (using Euclidean
distance) but also to compute the distance between subspaces. Some semantic fields
are rather closely related (e.g., sets of fruits and sets of vegetables), whereas other
fields are more distantly related (e.g., sets of animals and vehicles of conveyance).
In an analogy or metaphor, one relates two sets of two terms (which may be
stated either explicitly or implicitly). A good analogy is one in which the two sets of
terms are at correspondent points within their respective local subspaces. For exam¬
ple, “the Rolls-Royce is the lion of automobiles” will be a good metaphor to the
extent that one perceives the Rolls-Royce as the “king” of cars, as the lion is often
viewed as the “king” of the animals. A creative analogy is one in which the
between-subspace distance is large, but the corresponding within-subspace dis¬
tances are small. In other words, the two sets of two terms are at roughly correspon¬
dent points in their respective local subspaces, but the two sets of terms are in
relatively distant local subspaces. Thus, the correspondence between the two sets of
terms is good, but the two sets of terms are not trivially related. For example, “the
Rolls-Royce is the lion of automobiles” is a creative metaphor to the extent that
“Rolls-Royce” and “lion” are at correspondent points in their local subspaces, but
also to the extent that “automobiles” and “animals” are in relatively distant local
subspaces. The unusualness of the analogy or metaphor will be largely a function of
between-subspace distance. The accuracy of the analogy or metaphor will be largely
a function of alignment of corresponding points in their respective local subspaces.

Knowledge-acquisition components. Knowledge-acquisition components are in¬

volved particularly in a specialized form of creativity: insight. Janet Davidson and I
(Davidson & Sternberg, 1984; Sternberg & Davidson, 1982) have suggested that
three knowledge-acquisition components form the bases for three different kinds of
potentially creative insights.
1. Selective encoding. An insight of selective encoding involves sifting out rele¬
vant from irrelevant information. Significant problems generally present large
136 R. J. Sternberg

amounts of information, only some of which is relevant to problem solution. In¬

sights are involved in many fields of endeavor. For example, an insightful detective
needs to sift out relevant clues, and an insightful doctor needs to sift out relevant
symptoms in making a diagnosis. Davidson and I have found that a major dis-
tinguisher between better and lesser scientists is the ability of the better scientists to
recognize what is important in their data. The better scientists seem to know what
findings to emphasize and what findings to deemphasize or ignore. The lesser
scientists are less able to distinguish what matters more in their data and what
matters less. A famous example of a selective encoding insight in science can be
found in Sir Alexander Fleming’s discovery of penicillin. Fleming had been at¬
tempting to grow bacteria in a petri dish, but his culture was spoiled by a mold that
had killed the bacteria. A typical scientist might have despaired, his experiment
having been spoiled. Fleming noticed that what had started off as a mundane
experiment had become a monumental one: He had discovered a way of killing
bacteria via a mold, which later came to be called penicillin.
2. Selective combination. An insight of selective combination involves synthesiz¬
ing what might originally seem to be isolated pieces of information into a unified
whole that may or may not resemble its parts. Whereas selective encoding involves
knowing which pieces of information are relevant, selective combination involves
knowing how to blend together the pieces of relevant information.
There are many examples in real-world performance of how the insight of selec¬
tive combination operates. For example, once a detective has decided what clues are
relevant to the solution of a particular case, a way must be found to put them
together in order to pinpoint the perpetrator of the crime. A doctor must figure out
how the observed symptoms fit together in order to provide a basis for a diagnosis.
Also, we have found that the more competent scientists seem better able to fit their
findings together into a coherent package, or story. Less competent scientists are
less able to comprehend the interrelationships. They tend to present facts as iso¬
lated, rather than as part of a finely woven story relating to the phenomenon under
investigation. A famous instance of selective combination can be found in Darwin’s
formulation of the theory of evolution. The facts that compose the basis for this
theory had been available for a long time; what Darwin realized was how to put
these facts together to serve as the basis for a new theory of how life evolves.
3. Selective comparison. An insight of selective comparison involves relating
newly acquired information to information acquired in the past. Problem solving by
analogy, for example, is an instance of selective comparison: One realizes that new
information is similar to old information in certain ways (and dissimilar in other
ways) and uses this old information in better understanding the new information.
One could cite many instances of how selective comparison insights operate in
real-world activities. For example, a detective draws on experience in order to solve
a crime. The detective may recall a previous case in which the clues fit into a similar
pattern and use the earlier case as a basis for solving the newly committed crime. A
doctor may recollect a previous patient whose symptoms resembled those of a
Three-facet model of creativity 137

current patient. The diagnosis relevant in the earlier case might be relevant in the
current case. We have found that better scientists seem better able to relate their new
work to past endeavors in the field. They have a sense of where their work fits into
the current scientific picture, and they may carry on their work without a clear
understanding of why it is, or is not, really worth doing. A famous example of the
application of selective comparison is in Kekule’s discovery of the structure of the
benzene molecule. Kekule had been trying for some time to decipher the structure
of this molecule; one night he had a dream in which a snake danced around and bit
its tail. The next morning, Kekule made the association between the dream and his
quest: The snake biting its tail was a visual image for the structure of the ring.

The relation of intelligence to experience

The components of intelligence can be applied to problems at various levels of

experience, ranging from tasks and situations that are wholly unfamiliar to problems
and situations that are highly familiar, and possibly overleamed. Creativity can
come into play in both kinds of tasks and situations, either when an old and familiar
problem is seen in a new way or when a new and unfamiliar problem is seen in an
old way, that is, when one realizes how to apply what one already knows to a new
kind of task or situation.
Coping with novelty provides one of the best opportunities for display of
creativity. When one confronts something that is quite different from what is al¬
ready known, one has the opportunity to approach the problem in a particularly
creative way. In some of our research, for example, we present subjects with novel
kinds of tasks or situations, then ask the subjects to imagine what would follow if
some of these novelties were true. In one experiment, for example, subjects were
told to imagine themselves on the planet Kyron, where there are four kinds of
people: plins, who are bom young and die young; kwefs, who are bom old and die
old; baits, who are bom young and die old; and prosses, who are bom old and die
young (Sternberg, 1982; Tetewsky & Sternberg, 1986). In another experiment,
subjects were asked to solve analogies as though a counterfactual statement were
true, for example, that villains are lovable. We were interested in how well subjects
would reason in a world in which it is assumed that this statement is true (Marr &
Sternberg, 1986). Although almost everyone can solve problems of these kinds, we
have observed that some people can easily adjust to the demands of these non-
entrenched problems, whereas others have trouble imagining the world to be any
different from the way it is.
In recent times, cognitive scientists have more and more emphasized the role of
knowledge in intelligence (Chi, Glaser, & Rees, 1982), and knowledge would seem
to play a role in creativity as well. In particular, it is impossible to have novel ideas
about something if one knows nothing about it. One needs knowledge to extend
from in order to see how to apply or extend it creatively. At the same time, too
much knowledge can be (although need not be) a dangerous thing. One can become
138 R. J. Sternberg

so entrenched in set ways of seeing issues and problems that one is unable to go
beyond the existing paradigms and points of view. Thus, often the most creative
work in a given area is done by people who are relatively new to the field and who
know a fair amount about that field, but not too much about it.

The relation of intelligence to the external world

Intelligence has a contextual element, and so does creativity. According to the
triarchic theory, intelligence in the everyday environment can serve any one of three
contextual functions: adaptation to an existing environment, shaping of an existing
environment into a new environment, or selection of a new environment.
Creativity is probably most associated with environmental shaping. The great
creative thinkers in almost any field are those who shape the field to their own
conception of what that field should be. For example, a writer such as Hemingway,
an artist such as Picasso, an educator such as Dewey, and a scientist such as Herbert
Simon have all been viewed as creative in part because of their enormous impact in
shaping their respective fields. Sometimes, the shapers actually create new fields.
For example, Newton and Leibniz independently invented the field of calculus,
thereby shaping much mathematical thinking for centuries to come. Shaping an
environment so as to take a new form allows many possibilities for creativity,
although environments differ in the extent to which they can be shaped, and hence
in the extent to which one is able to exercise one’s creativity through shaping.
Environmental selection also enables one to exercise one’s creativity if one
moves into an unusual but nevertheless appropriate environment vis-a-vis one’s
abilities, propensities, and possibilities. This creativity can be exercised in even as
routine a decision as to where to go on a vacation. Some choices of vacations
(environmental selections) will turn out to be fairly routine, whereas others may be
much more exciting and creative.
Adaptation probably allows the fewest opportunities for true creativity, although
some adaptations to an existing environment can certainly be more creative than
others. For example, one may be able to stay within a given system because one
finds a way to work around that system, without actually changing the system.
Students in school often are able to adapt to the school environment precisely
because they find ways of working around it.
To summarize, a first facet of creativity is in the application of intelligence in a
creative - statistically unusual and highly appropriate - way. But there is more to
creativity than intelligence. A second facet of creativity derives from intellectual
style - the way in which intelligence is utilized in one’s thought and action.

Intellectual styles

A second aspect of creativity derives from the manner, or style, with which one
directs one’s intelligence. Some level of intelligence probably is a necessary condi-
Three-facet model of creativity 139

tion for creativity, but not a sufficient condition, in part because the style with
which one uses one’s intelligence is as important as the level of intelligence in
determining whether or not one is creative.
My comments regarding intellectual styles are drawn from my theory of mental
self-government as a basis for intellectual styles (Sternberg, 1986). The styles serve
to bridge between intelligence, on the one hand, and personality, on the other.
If one views intellectual functioning as mental self-government, then one can
view certain kinds of mental self-government as more potentially creative than
others. According to the theory, mental self-government can be characterized in
terms of five major aspects: (a) functions, which can be legislative, executive, or
judicial; (b) forms, which can be monarchic, hierarchic, oligarchic, or anarchic; (c)
levels, which can be global or local; (d) scope, which can be internal or external;
and (e) leaning, which can be conservative or progressive. Each of these aspects of
mental self-government has at least some relevance for understanding creativity.

Functions of government

Mental self-government, like societal government, serves three basic political func¬
tions: legislative, executive, and judicial. Consider each of these three styles in
turn.
The legislative style is found in people who (a) like to create their own rules, (b)
like to do things their own way, (c) prefer problems that are not prestructured or
prefabricated, (d) like to build structure as well as content, (e) prefer kinds of
activities that involve legislation, such as writing papers, designing projects, and
creating new business or educational systems, and (f) gravitate toward legislative
occupations such as creative writer, scientist, artist, sculptor, investment banker,
policymaker, or architect.
People with an executive style (a) like to follow rules, (b) like to figure out which
of existing ways to do things, (c) prefer problems that are prestructured or prefabri¬
cated, (d) like to fill in content within existing structure, (e) prefer activities that are
executive in nature, such as solving math problems, applying rules to problems,
giving talks or lessons based on others’ ideas, and enforcing rules, and (f) gravitate
toward executive kinds of occupations, such as lawyer, policeman, engineer,
builder (of others’ designs), surgeon, soldier, and proselytizer (of others’ systems).
Individuals with a judicial style (a) like to evaluate rules and procedures, (b) like
to evaluate existing structures, (c) prefer problems in which one analyzes and
evaluates existing things and ideas, (d) like to judge structure and content, (e) prefer
judicial kinds of activities such as writing critiques, giving opinions on things,
judging people and their work, and evaluating programs, and (f) prefer judicial
kinds of occupations such as judge, critic, program evaluator, admissions officer,
grant or contract monitor, systems analyst, or consultant.
I would argue that creative individuals tend to be those with a primarily legisla¬
tive style, or at least those who combine a legislative style with one of the two other
140 R. J. Sternberg

styles. Legislation is an activity that is creative by its nature. However, it is possible

to perform executive or judicial tasks in a way that emphasizes their legislative
functions. Note that a proclivity toward legislation does not guarantee an ability for
it. Whereas the preceding section of this chapter concerns the level of ability needed
for creative performance, this section deals with proclivity in the utilization of
abilities. And if one looks at creative performances in the world, these perfor¬
mances tend to be legislative, whether they are in creating art, literature, govern¬
mental legislation, scientific theories, or whatever. A creative administrator is one
who brings a legislative style to bear, at least in part, on administrative work,
whereas a “creative critic” is one who goes beyond just evaluative comments and
perhaps creates a new and exciting framework in which these comments are made.

Forms of mental self-government

Mental self-governments, like societal governments, come in multiple forms. Some

of these forms are more likely than others to lead to creative performance. The
monarchic style is shown by people who (a) tend to be motivated by a single goal or
need at a time, (b) are single-minded, (c) are driven, (d) believe that ends justify
means, (e) believe in going full speed ahead, and damn the obstacles, (f) are
oversimplifiers, (g) are un-self-aware, (h) tend toward intolerance, (i) are inflexi¬
ble, and (j) have little sense of priorities. People with the hierarchic style (a) tend to
be motivated by a hierarchy of goals, with the recognition that not all goals can be
fulfilled equally well and that some goals are more important than others, (b) tend
toward a balanced approach to problems, (c) believe that the ends do not justify the
means, (d) can deal with competing goals, (e) are complexifiers, (f) are self-aware,
(g) are tolerant, (h) are flexible, and (i) have a good sense of priorities. People with
the oligarchic style (a) tend to be motivated by multiple, often competing goals of
equal preceived importance, (b) follow multiple, possibly competing approaches to
problems, (c) are driven by goal conflict and tension, (d) believe that ends do not
justify means, (e) believe that competing goals and needs tend to interfere with task
completion, (f) are complexifiers, sometimes to the frustration point, (g) are self-
aware, (h) are tolerant, (i) are very flexible, and (j) have trouble setting priorities.
People with an anarchic style (a) are motivated by a potpourri of needs and goals
that are often difficult to sort out, (b) exhibit what seems like a random approach to
problems, (c) are driven by “muddle,” and sometimes seemingly inexplicable
forces, (d) believe that ends justify means, (e) are often unclear as to their goals, (f)
tend to be simplifiers, (g) are un-self-aware, (h) tend toward intolerance, and (i)
cannot set priorities because they have no basis for them.
Creativity probably is not as strongly linked to form of mental self-government as
it is to function. Nevertheless, one of the forms lends itself particularly to creative
potentialities, namely, the anarchic form. Anarchies have the ability to remove
themselves from existing constraints, ways of seeing things, and ways of doing
things. They are particularly drawn to problems that require a departure from
Three-facet model of creativity 141

existing rules and principles. Insight problems, for example, are often best solved
by people who can escape from existing mental sets and perceive problems in new
ways. But not all anarchies demonstrate creativity. Some of them lack the intellec¬
tual ability to be creative. Others lack the legislative style that needs to be combined
with the anarchic style in order to generate creative performance. Still others are
punished so strongly for their anarchic style that their creative ability is channeled
into antisocial pursuits. Anarchies generally are not to the tastes of either teachers or
parents, because the anarchies go against the existing grain. As a result, it is likely
that they will come into conflict with existing authority structures, and unless this
conflict is canalized in a constructive way, it may be defeating both to the anarchies
and for the society that ultimately could stand to benefit from the anarchies’ creative
performance.
True monarchies probably are the least likely to exhibit creative performance.
They are so susceptible to fixation on a single goal or means to a goal that they are
blinded to the alternatives. They simply do not seek or even envision alternative
ways of doing things. Creativity is possible within a hierarchical pursuit of a single
goal. But if both the means and the ends are monarchic, creativity is unlikely to
display itself.

Levels of mental self-government

Globalists are individuals who (a) prefer big issues, (b) ignore or eschew details, (c)
tend to be conceptualizers and idea-oriented, (d) enjoy abstract thinking, (e) show
an occasional or possibly frequent tendency to get lost on cloud 9, and (f) may see
the forest but overlook the trees. In contrast, localists (a) prefer details, (b) deal well
with details, (c) are oriented toward pragmatics, (d) tend toward concrete thinking,
(e) are down-to-earth, and (f) may miss the forest for the trees.
Most people are some combination of the global and local styles, and indeed the
two styles represent a continuum rather than a discrete bifurcation. Nevertheless, at
least some global inclination is needed for creative performance. Individuals who
get lost in details and who eschew the “big picture” are unlikely to put themselves
in a position where they can even imagine the creative possibilities of a problem
situation. They are too concrete and too detail-oriented. Global tendencies are
certainly not sufficient for creativity, but at least some sense for and desire to
achieve a big picture is necessary in order to approach a problem creatively.

Scope of mental self-government

Individuals with an internal style tend to be (a) introverted, (b) task-oriented, (c)
aloof, (d) socially less sensitive, (e) interpersonally less aware, and (f) alone when
they work. In contrast, those who emphasize the external scope of mental self-
government tend to be (a) extroverted, (b) people-oriented, (c) outgoing, (d) so¬
cially sensitive, (e) interpersonally aware, and (f) with others when they work.
142 R. J. Sternberg

Creativity is certainly possible within either internal or external scope of mental

self-government. However, at least some internal scope is needed, even in one’s
dealings with people, in order fully to display creativity. Creativity requires work¬
ing with the world of ideas, and this work is almost always, at least in part, a
solitary process. Some people gain from interactions with others, and they benefit if
they work well with others, but at some point along the way some solitary endeavor
is needed. To be creative, one needs to be able to be by and with oneself and also
needs to engage with ideas as well as things and others. People who are too external
in their orientation can be so busy seeking and enjoying the company of others that
they do not allow themselves the time and mental investment needed in order to
pursue creative endeavors.

Leaning of mental self-government

People with a conservative style (a) like to adhere to existing rules and procedures,
(b) like to minimize change, (c) prefer familiarity in life and work, and (d) avoid
ambiguous situations. In contrast, people with a more progressive style (a) like to
go beyond existing rules and procedures, (b) like to maximize change, (c) prefer
unfamiliarity in life and work, and (d) seek ambiguous situations.
Clearly, at least some progressivism is needed in order to display creative behav¬
ior. An individual who is conservative with respect to existing ways of seeing and
doing things is unlikely to look beyond them and hence is unlikely to find either
creative solutions or creative means toward problem solutions. The creative indi¬
vidual is one who expresses at least some dissatisfaction with existing principles and
procedures and directs this dissatisfaction in the constructive direction of seeking
new principles and procedures rather than merely criticizing or destroying old ones.
(Note that anarchies need not be progressives. Some anarchies may not care in what
directions things move so long as there is change, whereas true progressives wish
change in a forward direction and no other.)
To conclude, I have argued in this part of the chapter that the creative individual
not only needs certain intellectual abilities in order to be creative but also needs an
intellectual style that directs these abilities in a creative way. I have used my theory
of intellectual styles as mental self-government in order to argue this point. Certain
styles are more conducive than others to creative behavior. The most important
aspects of creative intellectual style is in a legislative bent toward problems. The
legislator is a person who enjoys looking for new things to do and for new ways of
doing old things. At least some progressivism in one’s approach to these legislative
problems facilitates the creative process, as does at least some internal orientation.
A creative person may approach problems from the standpoint of any of the various
forms of mental self-government, but true monarchies are unlikely to allow them¬
selves the mental and emotional freedom to be creative. In conclusion, creativity is
as much a function of the direction of one’s intelligence as it is of the level of one’s
intelligence. Both must be taken into account.
Three-facet model of creativity 143

Personality

It has been argued that two important facets of creativity are to be found in aspects
of intelligence and in aspects of the ways in which intelligence is utilized through
intellectual styles. A third and equally important ingredient of creativity is to be
found in personality. In particular, certain personality attributes are more conducive
to creative performance than are others.

Tolerance of ambiguity

As has been noted before (Vernon, 1970), tolerance of ambiguity is almost a sine
qua non of creative performance. Creative ideas rarely hatch in mature form.
Rather, as Gruber (1986) has pointed out, they evolve as systems. Although there
may be a few individuals who hatch their creative ideas whole, they are the excep¬
tion rather than the rule. The majority of creative people must learn to tolerate
ambiguity and incompleteness in the development of their creative products. The
development of this chapter, for whatever level of creativity it may involve, was an
example of this point. The idea for how to combine intelligence, intellectual styles,
and personality into a unified account of creativity evolved slowly, after many false
leads regarding how the chapter might be written. I found myself having to tolerate
ambiguity as I struggled for a synthesis that would make at least some contribution
to the literature on creativity. Had I not been willing to tolerate some ambiguity, the
chapter would never have been written. Similarly, in their work, scientists pursuing
a problem rarely solve it immediately, or with their first-pass attempt. Rather, their
solutions tend to be arrived at through a series of successive approximations, and
these successive approximations require them to know when they are closing in on a
solution but not quite there. Scientists who cannot tolerate this ambiguity may
declare themselves at a solution before they actually arrive. As a result, their
solution may be incomplete or simply wrong. Indeed, a constant struggle for many
scientists is the fear that their next idea or even their present idea will be their last.
They have to tolerate the ambiguity of not knowing if their creativity will continue
to manifest itself in future endeavors. The general point, then, is that to be creative,
one must be willing and able to tolerate at least some ambiguity in order fully to
manifest one’s creativity.

Willingness to surmount obstacles

Creative people inevitably encounter obstacles. So does everyone else. Such an

obstacle can be either an impediment to future creative performance or a spur
toward it. I have seen many a fine graduate student “go under” because of an
unwillingness or inability to persevere beyond the obstacles that always crop up in
graduate careers. Scientists have some articles rejected and grant proposals turned
down. Writers have manuscripts rejected or unfavorable reviews of some of their
144 R. J. Sternberg

work. Similarly, artists of various kinds also almost inevitably receive at least some
unfavorable reviews in the news media. Their ability to sustain their creativity
requires an ability to overcome these hurdles and not be dragged down by them. The
question is not whether or not an individual will encounter obstacles but rather how
one will handle them. A creative individual perseveres.

Willingness to grow

There exists a category of creative individuals in which one has one’s last and also
one’s major creative idea when one has one’s first major idea. These individuals are
unable or, more likely, unwilling to grow in their ideation. They come up with
creative ideas early in their careers, and then are afraid to move beyond those ideas.
Sometimes, perhaps, one of these individuals is overly reinforced for this first idea,
and continues to thrive on this reinforcement, all the more so when the reinforce¬
ment becomes intermittent, as it often does when one stays with a given idea for too
long a period of time. It is important to note that in Gruber’s theory, the systems of
creative people are evolving, not static. Those individuals whose systems are static
or show only minor evolution never achieve the true heights possible for creative
performance.

Intrinsic motivation

Amabile (1983) has pointed out the critical importance of intrinsic motivation for
creative performance. Ultimately, creative people receive much of their push from
their own internal desire to be creative. They motivate themselves. Individuals who
are largely extrinsically motivated are less likely to be creative, if only because they
are so dependent on reinforcement from others in the assessment of their work.
Hence, they are likely to be less willing to take the risks needed for creative
performance.

Moderate risk-taking

Creative people have a sense of acceptable levels of risk. In order to be creative,

they need to take some risk. At the same time, they need to recognize that there are
some projects, whatever the field, that are simply too risky to justify the investment
of their effort. Scientists, for example, often are aware of problems they would like
to study, but they do not tackle those problems because they are too difficult or too
unlikely to be susceptible to solution within a reasonable period of time. The
successful scientists choose the most difficult problems that they have a reasonable
expectation of being able to handle, not problems that are easier, and not ones that
are impossible for them.
Three-facet model of creativity 145

Desire for recognition

Creative individuals often balance high levels of intrinsic motivation with a desire
for recognition. Creativity does not exist in a vacuum; it is in part an ascription by
others. Hence, the creative individual will be identified as creative by virtue of
being so labeled by others. To achieve such a label, the person must seek it to at
least some extent and do what is necessary in order to achieve it.

Willingness to work for recognition

Some people wish to achieve recognition but are not willing to do what they need to
do in order to attain it. Ultimately, the individual recognized as creative not only
wishes to attain recognition but also actively seeks it in ways that are rewarded by
the groups making the relevant ascription.
To summarize, certain personality attributes tend to be associated with creative
performance. In this section of the chapter, I have described what some of these
personality attributes are. The list is not exhaustive, nor need a creative person
exhibit all these attributes. But at least some of them are likely to be present in
individuals recognized as creative.

Integration

People are creative by virtue of a combination of intellectual, stylistic, and person¬

ality attributes. This chapter has listed a subset of the attributes in each of these
three categories that contribute toward the generation and display of creative behav¬
ior. Moreover, it has outlined the scope of creativity as revealed by implicit theo¬
ries. In some domains, implicit theories probably are of little use: Fire, for example,
is what it is, regardless of what people think it is. The same cannot be said of
creativity: Because creativity is largely an attributional phenomenon, it is important
to know what people believe creativity is in order to know how they attribute
creativity to people.
Creativity can take many forms and come in many blends, in part because its
intellectual, stylistic, and personality facets can combine in essentially an infinite
number of ways. Some people will have more of the intellectual attributes, others
more of the stylistic attributes, and still others more of the personality attributes. As
a professor, for example, I am aware of students with creative intellectual abilities
who simply do not have the stylistic or personality attributes for their creativity to
come through; other students have the right personality attributes, but frustrate
others and usually themselves by not having the intellectual wherewithal to produce
creative products. The general point is that creativity will manifest itself in different
forms depending on the blend of characteristics one brings to one’s attempts at
creative performance.
146 R. J. Sternberg

Different blends result not only from individual differences across the three facets
of the model of creativity but also from different blends within each facet. Some
individuals, for example, may have something of the anarchic style that would free
them from past constraints, but may be so local in their anarchistic tendencies, or so
executive rather than legislative (e.g., the individual who carries out the revolution
that others plan) in their functioning, that their tendency toward anarchism does not
manifest itself at all in creative behavior. In order fully to understand the mental
bases of creativity, we need to understand both the interrelations and the intrarela¬
tions within each of its facets.
Even this is not enough, though. It seems likely that certain blends are more
likely to be synergistic (and other blends antagonistic) with respect to generating
creative behavior. It does not seem likely that the parts are merely additive, so that
the whole represents a simple sum of its parts. For example, a legislative style
would seem to be particularly synergistic with excellence in the execution of the
legislative metacomponents: problem recognition, problem definition, component
selection, strategy selection, and representation selection. In contrast, an executive
style combined with excellence in these metacomponents might be antagonistic to
the display of creativity: The legislative metacomponents might well end up being
used in the service of executive functions.
The notion of “service” of one aspect of mental functioning to another is one
that I believe is particularly important in the display of creativity. All styles repre¬
sent a continuum rather than a discrete partition, and no one is completely dedicated
to one style or another: In order to function in the world, they could not be. As a
result, people will display different styles as a function of their predispositions, as
well as of the tasks and situations in which these predispositions function. Someone
who is primarily executive may use legislative activities in the service of the
executive, or vice versa. Someone who has a great desire for recognition may seek
to turn even executive tasks into legislative ones that are likely to lead to increased
recognition. In seeking to understand creativity, therefore, we need to understand
the interactivity among its parts as well as their independent functioning.
The emphasis in this chapter has been on the internal attributes that are likely to
lead to creative functioning. But a complete model of creativity would have to take
into account the environmental as well as the personal variables that facilitate as
well as impede the manifestation of creativity. A potentially creative individual may
wither in an environment that does not foster, or that actively inhibits, a display of
creative behavior. Such environments, unfortunately, are not rare. I have argued
elsewhere (Sternberg, 1986) that certain types of schooling, for example, can inhib¬
it creativity, and there are certainly job settings in which creative displays are
punished rather than rewarded. A complete theory of creativity, therefore, will
ultimately be as much a theory of environments as of persons, and some of the other
chapters in this volume deal with the environmental variables that tend either to
foster or to impede a display of creativity.
In conclusion, creativity is a multifaceted phenomenon, of which three critical
Three-facet model of creativity 147

facets would seem to be aspects of intelligence, style, and personality. This chapter
has described what some of these aspects might be and how they might be utilized in
ways that render people more or less creative in their thought and action.

References
Amabile, T. M. (1983). The social psychology of creativity. New York: Springer-Verlag.
Chi, M. T. H., Glaser, R., & Rees, E. (1982). Expertise in problem solving. In R. J. Sternberg (Ed.),
Advances in the psychology of human intelligence (Vol. 1, pp. 7-75). Hillsdale, NJ: Lawrence
Erlbaum.
Collins, A. M., & Quillian. M. R. (1969). Retrieval time from semantic memory. Journal of Verbal
Learning and Verbal Behavior, 8, 240-248.
Davidson, J. E., & Sternberg, R. J. (1984). The role of insight in intellectual giftedness. Gifted Child
Quarterly, 28, 58-64.
Feldman, R. D. (1982). Whatever happened to the quiz kids? Chicago: Chicago Review Press.
Getzels, J., & Csikszentmihalyi, M. (1976). The creative vision: A longitudinal study of problem-finding
in art. New York: Wiley-Interscience.
Gruber, H. E. (1986). The self-construction of the extraordinary. In R. J. Sternberg & J. E. Davidson
(Eds.), Conceptions of giftedness. Cambridge University Press.
Marr, D. B., & Sternberg, R. J. (1986). Analogical reasoning with novel concepts: Differential attention
of intellectually gifted and nongifted children to relevant and irrelevant novel stimuli. Cognitive
Development, 1, 53-72.
Palmer, S. E. (1975). Visual perception and world knowledge: Notes on a model of sensory-cognitive
interaction. In D. A. Norman & D. E. Rumelhart (Eds.), Explorations in cognition. San Francisco:
Freeman.
Perkins, D. (1981). The mind’s best work. Cambridge, MA: Harvard University Press.
Shepard, R. N., & Metzler, J. (1971). Mental rotation of three-dimensional objects. Science, 171, 701-
703.
Sternberg, R. J. (1982). Natural, unnatural, and supernatural concepts. Cognitive Psychology, 14, 451-
458.
Sternberg, R. J. (1985a). Beyond IQ: A triarchic theory of human intelligence. Cambridge University
Press.
Sternberg, R. J. (1985b). Implicit theories of intelligence, creativity, and wisdom. Journal of Person¬
ality and Social Psychology, 49, 607-627.
Sternberg, R. J. (in press). Mental self-government: A theory of intellectual styles and their develop¬
ment. Human Development.
Sternberg, R. J., & Barnes, M. (1985). Real and ideal others in romantic relationships: Is four a crowd?
Journal of Personality and Social Psychology, 49, 1589-1596.
Sternberg, R. J., & Bower, G. H. (1974). Transfer in part-whole and whole-part free recall: A com¬
parative evaluation of theories. Journal of Verbal Learning and Verbal Behavior, 13, 1-26.
Sternberg, R. J., Conway, B. E., Ketron, J. L., & Bernstein, M. (1981). People’s conceptions of
intelligence. Journal of Personality and Social Psychology, 41, 37-55.
Sternberg, R. J., & Davidson, J. E. (1982, June). The mind of the puzzler. Psychology Today, 16, 37-
44.
Tetewsky, S. J., & Sternberg, R. J. (1986). Conceptual and lexical determinants of nonentrenched
thinking. Journal of Memory and Language, 25, 202-225.
Tourangeau, R., & Sternberg, R. J. (1981). Aptness in metaphor. Cognitive Psychology, 13, 27-55.
Tulving, E. (1966). Subjective organization and effects of repetition in multi-trial free-recall learning.
Journal of Verbal Learning and Behavior, 5, 193-197.
Vernon, P. E. (1970). Creativity. Harmondsworth, UK: Methuen.
Zuckerman, H. (1983). The scientific elite: Nobel laureates’ mutual influences. In R. S. Albert (Ed.),
Genius and eminence: The social psychology of creativity and exceptional achievement (Vol. 5, pp.
241-252). New York: Pergamon Press.
6 Problem solving and creativity

Robert W. Weisberg

It has long been believed by psychologists that research in problem solving will
illuminate broad issues in the study of creative thinking, defined as the thought
processes involved in producing work of acknowledged greatness in art or in sci¬
ence. The gestalt psychologists, for example, explicitly equated the productive
thinking involved in true problem solving with the thought processes occurring in
more esoteric situations, such as the work of great scientists (Wertheimer, 1959).
However, the direct connection between research on problem solving and study of
creative thinking has been slow in coming. Modem cognitive psychologists have
only very recently begun to extend conclusions drawn from laboratory research in
problem solving to the thought processes used by scientists and artists (Greeno,
1980; Perkins, 1981; Weisberg, 1986). As one example of this separation of the
problem-solving literature from that on creativity, Newell and Simon’s (1972)
influential analysis of problem solving does not even have entries for creativity or
creative thinking in the index. Thus, even though Newell and Simon had proposed
an analysis of creative thinking in earlier publications (Simon, 1979), the relevance
of the work on problem solving to an understanding of creativity was at most
implicit in their magnum opus. As a further example of this conceptual bifurcation,
examination of several modem texts on cognitive processes (Bransford, 1979;
Gilhooly, 1982; Martindale, 1981) indicates a lack of connection between discus¬
sions of problem solving and creative thinking, if both are even mentioned in the
same work. For the most part, then, cognitive psychologists have devoted their
efforts to the study of laboratory problems of various sorts and seem to have
assumed that creative thinking would ultimately be illuminated by this work, with¬
out explicitly dealing with creative thinking per se.
Perhaps in part because of this lack of direct involvement by experimental psy¬
chologists, the literature on creativity was until recently dominated by what one
could call the “genius” view of creativity, which also pervades our society. This
view, which has many sources, ranging from Plato to Koestler (1964) to Osborn
(1953) to psychometric theorists such as Guilford (1950), assumes that truly cre¬
ative acts involve extraordinary individuals carrying out extraordinary thought pro¬
cesses. These individuals are called geniuses, and the psychological characteristics

148
Problem solving and creativity 149

they possess - cognitive and personality characteristics - make up what is called

genius.
Some support for the genius view came from self-reports by acknowledged
creative geniuses, such as the following from Samuel Taylor Coleridge, which
describes the circumstances leading to his writing “Kubla Khan.” At the time,
Coleridge had been ill and was living in the country. One afternoon, because of ‘‘a
slight indisposition,” he took “an anodyne” (opium) and fell asleep while reading
a book describing Kubla Khan’s capital city. According to Coleridge (Ghiselin,
1952, pp. 84-85), the following events then occurred:
The Author continued for about three hours in a profound sleep, at least of the external
senses, during which time he had the most vivid confidence that he could not have composed
less than from two to three hundred lines; if that indeed can be called composition in which all
the images rose up before him as things, with a parallel production of the concurrent
expressions, without any sensation or consciousness of effort. On awakening he appeared to
himself to have a distinct recollection of the whole, and taking his pen, ink, and paper,
instantly and eagerly wrote down the lines that are here preserved. At this moment he was
unfortunately called out by a person on business from Porlock, and detained by him above an
hour, and on his return to his room, found, to his no small surprise and mortification, that
though he still retained some vague and dim recollection of the general purport of the vision,
yet, with the exception of some eight or ten lines and images, all the rest had passed away like
the images on the surface of a stream into which a stone has been cast, but alas! without the
after restoration of the latter!

Thus, an entire poem appeared full-blown, with no conscious work on Cole¬

ridge’s part. This example has been taken as evidence for the importance of uncon¬
scious processes in creativity. A similar report was made by Mozart in an often-
quoted letter (Ghiselin, 1952, pp. 44-45):
When I feel well and in a good humor, or when I am taking a drive or walking after a good
meal, or in the night when I cannot sleep, thoughts crowd into my mind as easily as you could
wish. Whence and how do they come? I do not know and I have nothing to do with it. Those
which please me, I keep in my head and hum them; at least others have told me that I do so.
Once I have my theme, another melody comes, linking itself to the first one, in accordance to
the needs of the composition as a whole.

Mozart’s report may be even more remarkable than that of Coleridge, because
Mozart seems to have needed no help from stimulants or other external sources in
order to have creative products come to him. Again, unconscious thought is as¬
sumed to have been crucial here.
A final famous example of the extraordinary nature of creative thinking is
Friedrich August von Kekule’s report of his discovery of the ring structure of
benzene. According to Kekule, he had the following dream during days of thinking
about the problem (Rothenberg, 1979, pp. 395-396):

I turned my chair to the fire and dozed. Again the atoms were gamboling before my eyes.
This time the smaller groups kept modestly to the background. My mental eye, rendered
more acute by repeated vision of this kind, could now distinguish larger structures, of
manifold conformation; long rows, sometimes more closely fitted together; all twining and
150 R. W. Weisberg

twisting in snakelike motion. But look! What was that? One of the snakes had seized hold of
its own tail, and the form whirled mockingly before my eyes. As if by a flash of lightning I
awoke. . . . Let us learn to dream, gentlemen.

In discussions of this incident (Koestler, 1964), two points are emphasized. First,
Kekule was dreaming, which points again to the difference between creative think¬
ing and ordinary conscious thinking. Second, Kekule used a visual analogy in order
to solve his problem. He represented the strings of atoms as snakes, and then one of
them bit its tail. Thus, the use of a remote analogy (snakes are far removed from
atoms) points to still another difference between creative thinking and ordinary
conscious thinking. Many theorists have emphasized the role of analogy in creative
thought (Perkins, 1981). Another often-cited example is Bohr’s use of the solar
system as a model for the structure of the atom. The genius view emphasizes the
fact that these two cases involved remote analogies (i.e., the domains involved in
each example are not ordinarily considered to be closely related). Strings of atoms
do not seem to be closely related to snakes, and neither do atoms and solar systems
seem closely related, but creative thinkers found these remote analogies useful.
Thus, according to the genius view, the creative thinker finds connections where
ordinary individuals do not (Koestler, 1964).
In sum, these examples make clear the purportedly extraordinary aspects of great
creative acts and the individuals who produce them. Mozart reported that melodies
came to him unbidden, and by themselves expanded into complete works, without
his conscious participation. Coleridge reported a similar experience, helped along
by a dose of opium. Finally, Kekule used a visual analogy in a dream to solve the
problem of the structure of benzene. In these examples we see unconscious thought
processes, as well as modes of conscious thought (e.g., use of a remote analogy),
that are beyond the reach of ordinary individuals.
The genius view of creativity has motivated students of creativity to examine
those characteristics that lead to the production of great artistic and scientific works
(Weisberg, 1986). One stream of research has focused on isolating those personality
characteristics that supposedly separate the creative genius from ordinary indi¬
viduals. There have also been examinations of the brains of geniuses, such as
Einstein, in order to determine the areas of the brain that might underlie the capacity
to be creative. Finally, much research has been concerned with the mode of thinking
involved in creativity. The genius view assumes that creative thinking involves
some kind of thought processes (and related personality characteristics) that allow
one to break away from the habitual, away from one’s past experience, and thereby
produce something truly novel, in a leap of insight. As was seen in the foregoing
quotations, unconscious processes, altered states of conscious, and use of remote
analogies are three characteristics that have been attributed to creative thinkers.
My purpose in this chapter is to explore the implications of laboratory research in
problem solving for an understanding of significant creative acts, and thereby to
raise questions about the genius view of creativity. The analysis will begin with a
consideration of several potentially significant findings from laboratory study of
Problem solving and creativity 151

undergraduates’ performances on a problem-solving task. These findings will then

be supported by results from other laboratory investigations of problem solving.
This will lead to a discussion of several case studies from the literature on creativity
in science, technology, and the arts. Although the data from such case studies are
not as rigorous as those from laboratory experiments, it is my belief that one must
directly address the literature concerning individuals of acknowledged greatness. If
not, the question always remains whether or not one can generalize from under¬
graduates solving problems in an experiment to, say, Darwin, Watson and Crick,
Edison, Picasso, Mozart, and so on. In making any such generalization, the burden
is on the one who would make the generalization from undergraduates to creative
geniuses, and so I shall make clear what I see as the connection between laboratory
work in problem solving and specific events in the careers of great artists and
scientists. The conclusions to be drawn may not at first glance appear to be relevant
to true creative thinking, given the genius view. I shall take those conclusions
seriously, however, in order to show that the thought processes underlying problem
solving by ordinary undergraduates may be similar to those involved in much more
esoteric activities, such as developing a scientific theory or painting a mural.

Solution of the candle problem

Duncker’s (1945) candle problem asks that the problem solver attach a candle to the
wall, with the available materials being a book of matches and a box of tacks or
nails. This problem was originally of interest to gestalt psychologists because it was
assumed to be an example of a situation in which past experience interfered with
productive thinking, in a manner supposedly comparable to that seen in the Luchins
water-jar experiments (Luchins & Luchins, 1959). It was argued by Duncker and
others that one very effective but infrequent solution to this problem (using the tack
box as a platform or holder for the candle) was infrequent because subjects were
“fixated” on the use of the box as a container for the fasteners and therefore were
not able to conceive of the box as a platform for the candle.
Research by Weisberg and Suls (1973) pointed to a more straightforward expla¬
nation of why subjects produced or failed to produce the box solution. Lirst, com¬
parisons of thinking-aloud verbal protocols produced by box-solvers with those of
non-box-solvers indicated that all individuals began to work by attempting to apply
their knowledge relatively directly to the problem, in response to the instructions
(Weisberg & Suls, 1973, Exp. 3). Thus, all individuals started by trying to follow
the problem instructions as straightforwardly as they could - by attempting to attach
the candle to the wall using the fasteners or melted wax as glue. If these initial
attempts were found to be inadequate in some way (e.g., the candle started to split
when the fastener was pushed into it; the candle was too thick for the fastener to
pass through; or the wax-glue was not strong enough), then attempts would be made
to correct this inadequacy. Other solutions, including the box solution, evolved out
of these attempts to correct inadequacies in seemingly straightforward solutions.
152 R. W. Weisberg

This research indicated that creative solutions to problems were based on interaction
between the subject’s knowledge and the problem itself. The subject sets the criteria
for solution, based on the instructions and knowledge, and then uses these criteria as
the basis for judging the adequacy of any attempted solution. Inadequacies in any
solution then become further problems to solve, and so on, recursively, until all
criteria are met.
This solution process could be conceptualized as multiple searches of memory, in
the sense that each step in the solution, if judged to be inadequate, serves as the cue
for further retrieval of information. One could say that the problem becomes refor¬
mulated, or changed into a new problem, as the problem solver works through it.
New information constantly becomes available as the subject realizes that what was
conceived of as an adequate solution is inadequate in some way. This recursive
search is similar to Newell and Simon’s (1972) discussion of means-ends analysis
as a problem-solving heuristic. In contrast to the genius view, which postulates
leaps of insight based on remote analogies, this recursive memory search is “local”
in nature, because it depends at any time only on the inadequacy the individual is
attempting to deal with.
It must be emphasized, however, that this process is more than simply retrieval
and use of some old knowledge, because the problem situation is new to our
subjects. Therefore, by definition, they cannot be simply retrieving something
previously done in this situation. Any retrieval that occurs must be brought about
through only a partial match between the present situation and past experience. In
this sense, then, all solutions to the candle problem (and perhaps all solutions to any
problem) should be considered creative, so long as they are novel as far as the
problem solver is concerned, and they solve the problem at hand. I shall therefore
assume that the various solutions to the candle problem are equivalent psychologi¬
cally; production of one solution rather than another depends on the particular tack
taken by a given problem solver, which in turn depends on those specific factors in
the problem that the problem solver considers important. One subject might respond
to the weakness of the wax-glue by trying to increase the area of contact between the
candle and the wall, whereas another might try to provide additional support for the
glue. The former subject might try to flatten the candle, whereas the latter might try
to use the box as a shelf under the candle. Although earlier investigators emphasized
lack of production of the box solution, from the present perspective all the various
solutions are equivalent, because they are responses to a novel situation, they meet
the criteria, and, I believe, they are the results of the same thought processes.

Implications

This work has several implications for understanding problem solving in particular
and creative thinking in general. The first implication is that all attempts to solve
problems are firmly based on past experience. Individuals initiate work on a prob¬
lem based on a match between that problem and their knowledge. Furthermore, any
Problem solving and creativity 153

inadequacies discovered in an initial solution are also dealt with on the basis of past
experience (Greeno, 1980). Second, novel solutions to problems come about in an
evolution, as one gradually moves away from the conception with which one began,
through local memory searches. No great leaps of insight occur in the production of
any solution, whether or not that solution is judged by the experimenter to be
particularly creative. However, in order to see the incremental nature of problem
solving, one must have available more than simply the initial problem and the final
solution. That is, if one knew only that a subject had produced the box solution to
the candle problem, one might be tempted to assume that that solution had come
about in a modest leap of creativity. With a detailed problem-solving protocol
available, on the other hand, one might see that the subject began with an attempt to
glue the candle to the wall with melted wax. The box was then used in an attempt to
shore up the wax. Thus, the incremental nature of creative problem solving is
revealed by a fine-grained analysis of the situation. Third, as just mentioned, this
incremental process based on local memory search is set in motion by feedback
concerning the inadequacy of some proposed solution. This new information trans¬
forms the problem situation and thereby makes possible new solution types. To
continue with the same example, a subject who finds that the wax-glue is not strong
enough to keep the candle affixed to the wall is now faced with a new problem: How
does one hold up a falling candle if glue does not work? This new problem can
result in an evolution of the initial solution into something new. Fourth, if it is true,
as mentioned earlier, that all solutions to problems are “creative,” so long as they
are novel and they meet the demands of the problem, then the capacity to think
creatively must be a basic human capacity, and not the exotic trait or skill envi¬
sioned by the “genius” view. This has potentially important implications for teach¬
ing creative thinking.

Research with other problems

These implications, though evolving out of work on a single laboratory problem,

have support from other laboratory research in problem solving.

Dependence of creative problem solving on past experience

There has been a small but steady stream of research concerning what are called
“insight problems.” This is a heterogeneous class of problems that are assumed to
be difficult to solve because subjects are inhibited by past experience. Supposedly,
the only way to solve such problems is to break away from past experience, which
allows one to approach the problem in a novel way (Adams, 1979; Scheerer, 1963).
An example of such a problem is the nine-dot problem, shown in Figure 6.1.
According to the traditional analysis, the nine-dot problem would be very easy if
subjects could simply break away from their tendency to keep their lines within the
square formed by the dots. Once subjects break out of this fixation on the square
154 R. W. Weisberg

Figure 6.1. Nine-dot problem.

shape, and thereby approach the problem on its own terms, solution should be
effortless.
Two studies (Burnham & Davis, 1969; Weisberg & Alba, 1981) tested this view
by instructing subjects that in order to solve the nine-dot problem, they had to draw
their lines outside the square formed by the dots. Contrary to the traditional view, in
both studies only a small percentage of subjects solved the problem, and those
subjects took a relatively long time to do so. In addition, both studies found that
subjects needed relatively specific information in order to solve the problem.
Weisberg and Alba were also able to facilitate or inhibit solution of the nine-dot
problem by giving subjects experience solving dot problems of different sorts.
These results were extended by Lung and Dominowski (1985), who facilitated
solution of the nine-dot problem by giving subjects practice with six simple dot
problems that required that the lines be drawn outside the dots, as well as by giving
subjects more detailed instructions concerning what they had to do once they drew
lines outside the square formed by the dots. This study supports the earlier ones in
showing that the nine-dot problem demands relatively extensive knowledge con¬
cerning dot problems on the part of the problem solver.
Similar conclusions can be derived from recent research on expert problem solv¬
ing in a variety of domains. This work, which began with DeGroot’s (1966) analy¬
sis of problem-solving skills of master chess players, has resulted in the conclusion
that master-level problem solving in any domain depends on the acquisition of deep
knowledge about the domain in question (Greeno, 1980). Studies that have com¬
pared problem-solving performances of experts and novices in such domains as
computer programming, mechanics, and geometry have all indicated that experts
are able, because of their knowledge, to “home in” on the important aspect of any
novel problem. The expert is able to relate a novel problem to something already
known and to use this knowledge as the basis for dealing with the new problem. As
Greeno (1980) notes, there has been a blurring of the distinction between “real”
problem solving and behavior based on “mere knowledge,” as it has been realized
that all problem solving is based on knowledge. Any differences between doing
Problem solving and creativity 155

something by rote versus as a result of creative thinking may be matters of degree

rather than of kind.
Thus, research indicates that rather than being independent of past experience,
truly efficient problem solving comes about only when an individual has acquired a
deep knowledge of the domain in question. As mentioned earlier, it may be reason¬
able to assume that all problem solving, if the problem is novel, is creative. This in
turn leads to the speculation that creative thinking in other domains, such as science
and the arts, may also depend on extensive knowledge in those domains. Evidence
relevant to this speculation will be considered in a later section.

Stimuli to novelty

It was mentioned earlier that evidence indicates that the candle problem can become
changed as the individual works through it. More specifically, as the problem solver
realizes that an initial solution will not work, this information serves as the basis for
a new search of memory, in an attempt to overcome the difficulty. Novel solutions
to a given problem thus come about because the new information moves the indi¬
vidual away from the original conception. It should therefore be possible to isolate
the information that pushes the problem solver in a given direction and use it to
induce naive problem solvers to approach the problem in that way. That is, it ought
to be possible to short-circuit the incremental processes that lead to innovation in
problem solving by presenting the critical information at the very beginning. This
hypothesis was tested by Weisberg and DiCamillo (1986) in a series of experiments
using what we call the Charlie problem:

Dan came home and found Charlie dead on the floor and Tom in the same room. On the floor
Dan saw some broken glass and water. How did Charlie die?

This problem was studied by having subjects ask yes/no questions of the experi¬
menter, which enabled us to keep track of the information they acquired as they
worked. (The solution: Tom [a cat] knocked over a fishbowl. Charlie [a fish] died
from lack of oxygen.)
When subjects are presented with this problem, all initially assume that the
characters are human and that the pieces of glass and the water came from a broken
drinking glass. These initial assumptions follow directly from the way the problem
is presented. The characters have human names, and no information is given that
would cause the subject to question whether or not they are in fact human. In order
to solve the problem, however, the subject must come to the realization that the
characters are not human, and so on. The research indicated that the solution to the
problem only gradually developed, as subjects acquired information in response to
their yes/no questions.
As in the candle problem, all subjects begin the Charlie problem in the same way:
They ask questions about human deaths. Eventually subjects acquire the informa¬
tion that Charlie died from lack of oxygen. This information alone is not enough to
156 R. W. Weisberg

push them to the realization that Charlie is a fish. For about half our subjects, once
they found out that Charlie died from lack of oxygen and that the glass and water
came from a broken fishbowl, they then inferred rather directly that Charlie was a
fish. Thus, subjects’ novel analysis of the situation seemed to be the result of their
acquisition of two critical pieces of information that changed the situation they were
dealing with.
The other subjects did not go directly from “Charlie is human” to “Charlie is a
fish.” These individuals first made an intermediate deduction, that Charlie was not
human, which then led them to “Charlie is a fish.” The switch to “Charlie is not
human” was stimulated by new information that was inconsistent with the initial
assumption that Charlie was human. As an example, one subject asked about
Charlie’s age and found that though he might be a year old, he was not an infant.
That information led rather directly to the conclusion that Charlie was not human.
Thus, two modes of solution were found, one going directly from “Charlie is
human” to “Charlie is a fish,” and one going indirectly, through “Charlie is not
human.” Furthermore, different sorts of information seemed to trigger the two
modes. This was verified in later experiments in which naive subjects who were
given the supposedly relevant information not only solved the problem more quick¬
ly but also solved it using the expected mode of solution.
In sum, research using the Charlie problem has demonstrated that novel solutions
evolve out of the relatively prosaic solutions with which all subjects begin, in this
case proposals concerning human deaths. Furthermore, it has been possible to
isolate the crucial pieces of information that result in formulation of the novel
solution. Thus, in this case at least, novelty in problem solving comes about through
ordinary thought processes in interaction with the information available in the
problem. As this information changes, so does the solution produced by the subject,
but the thought processes remain the same.

Problem solving, predictability, and insight

Metcalfe (1986) has recently presented evidence to support the view that there is a
mode of problem solving that does not depend on past experience and that should be
characterized as resulting from leaps of insight. Metcalfe has asked subjects to
predict how well they will perform on a task, before they begin to work on it. When
the task involves answering trivia questions, the subjects are able to predict which
questions they will be able to answer. However, when the task involves problem
solving, there are cases in which subjects cannot predict how they will perform.
Thus, in tasks involving retrieval of information from memory, subjects are able to
predict their performance, whereas they are not always able to do so in tasks
involving problem solving.
Metcalfe takes this discrepancy as evidence for the view that in some cases
problem solving does not depend on retrieval of information from memory, contrary
to the view presented here. According to Metcalfe, because subjects can predict
Problem solving and creativity 157

how well they will be able to retrieve information from memory, then if problem
solving were based on retrieval of information from memory, they should also be
able to predict how well they will solve problems.
However, Metcalfe’s analysis is based on an oversimplification of the processes
involved in problem solving. Whereas the present view argues that problem solving
always depends on the use of past experience, and there is no need to postulate a
special process called insight, this does not mean that subjects must be able to
predict their problem-solving performance. Indeed, most of the “insight” problems
studied by psychologists are so designed that prediction of performance is essen¬
tially impossible (Weisberg, 1980, chaps. 10 and 12), which was what made such
problems interesting in the first place. Problems like the Charlie problem are so
designed that the initial direction taken by the subject is incorrect, which means that
any initial predictions by naive subjects are certain to be unrelated to final perfor¬
mance. Solution of such problems depends on feedback from the environment,
which informs the problem solver that the initial attempt is incorrect. Therefore,
Metcalfe’s finding that subjects cannot predict their performance is not surprising.
From my perspective, even though problem solving is firmly rooted in past experi¬
ence, this does not mean that problem solving will be predictable on the part of the
subject.

“Local” memory search

A number of studies support the claim that individuals do not widely search memory
when attempting to solve problems. The general theme of this research is to provide
problem solvers with information demonstrably relevant to the problem at hand, and
then to determine the factors that influence the utilization of that information. Over
the years there have been studies of the effectiveness of hints of various sorts during
problem solving; see Bourne, Ekstrand, and Dominowski (1971) for a review. I
shall concentrate on more recent research.
In one study (Weisberg, DiCamillo, & Phillips, 1978), subjects received paired-
associate training before trying to solve the candle problem. For the experimental
subjects, the critical pair was “candle-box,” which was presumably relevant to the
candle problem. Control subjects learned “candle-paper” in place of the critical
pair. It was found that without an explicit hint that the previously learned pairs
might help them solve the problem, experimental subjects produced box solutions
no more frequently than did control subjects. This was true even when actual
objects, identical with those in the candle problem, were used in the paired-associ¬
ated training. In this condition, learning the candle-box pair involved the subjects
actually putting the candle into the box, but facilitation in using the box to solve the
candle problem still was not achieved. On the other hand, all subjects given a hint
that one of the pairs might help them solve the problem did produce box solutions,
indicating that the candle-box association was potentially useful. In explaining these
results, it seems that any memory search stimulated by the candle problem was very
158 R. W. Weisberg

limited in nature and was restricted to previous situations involving attaching things
to walls. If so, then without an explicit hint to refer to the paired associates, one
would not expect them to help, even though the experimenter, from outside the
situation, knows that one pair is potentially relevant.
This conclusion is further strengthened by results of studies using verbal prob¬
lems of the following sort (Perfetto, Bransford, & Franks, 1983):

A man in a certain town has married 20 women in the town. The women are all still alive.
The man is neither a Mormon nor a bigamist, he is not divorced, and he has not broken any
law. How is that possible?

A police officer watched an individual go against a red light, ignore a stop sign, and go the
wrong way down a one-way street, and yet the officer did nothing. Why not?

The solution to the first problem is that the man in question is the clergyman who
performed the service marrying each of the 20 women to her husband. In the second
problem, the police officer did nothing because the individual in question was a
pedestrian.
In their research, Perfetto and associates first presented subjects with sentences to
study, ostensibly as part of a memory task. Several of the sentences were potentially
relevant to the word problems presented in the second part of the experiment. An
example: A minister may marry several people in a single week. Although such a
sentence seems of obvious relevance to the problem presented earlier, Perfetto and
associates, like Weisberg and associates (1978), found no effect of previous ex¬
posure to the critical sentences, unless subjects were explicitly told that some of the
earlier-studied sentences might be relevant to the problems.
The local nature of memory search during problem solving has also been shown
in recent studies investigating the role of analogy in problem solving. As mentioned
earlier, one component of the genius view is that creative products sometimes result
from leaps of insight based on remote analogies. In one set of studies on the role of
analogy in problem solving, Gick and Holyoak (1983) used Duncker’s (1945)
radiation or tumor problem as the target problem. In this problem, a doctor is faced
with a patient with an inoperable stomach tumor and has available a source of rays
that at sufficient intensity will destroy it. However, the rays at that intensity will
also destroy healthy tissue, which is unacceptable. How can the doctor use the rays
to destroy the tumor while not destroying any healthy tissue? The solution is to use
two or more weak bundles of rays, which are so positioned that they cross just at the
tumor. The intensity of the rays at the tumor will then be sufficient to destroy it,
while the lower-intensity individual bundles will not destroy healthy tissue.
Gick and Holyoak gave subjects prior experience with stories analogous to the
radiation problem and then compared their performance to that of control subjects
who had preexposure to nonanalogous stories. An example of an analogous story
used by Gick and Holyoak involves a general who wishes to attack a fortress, which
has roads radiating out from it like the spokes of a wheel. The general has his army
massed, ready to attack the fortress, but the roads leading to it have been mined, so
Problem solving and creativity 159

that if the whole army is sent down any one road, the weight of soldiers and
machines will set off land mines, causing destruction of the army as well as wide¬
spread damage to houses of citizens who live along the road. The general therefore
divides his army into small groups and positions each group at the head of one of the
roads leading to the fortress. At a signal from the general, the small groups move
together toward the fortress, and because each group is too small to set off the
mines, they are all able to reach the fortress at the same time, and so the general can
have his entire army attacking the fortress at once.
Gick and Holyoak (1983) found that preexposure to two stories like the general
story facilitated performance on the radiation problem. Most important in the pre¬
sent context, this facilitation was found even among subjects who were not in¬
formed of the relationship between the stories and the radiation problem. These
results support the view that creative problem solving can use remote analogous
relations as the basis for memory search.
However, there is one potential contaminating factor in Gick and Holyoak’s
studies that raises questions about how wide a search of memory was actually made
by their subjects. In addition to the analogous structural relations that existed
between the stories and radiation problem, they were also linked by context: They
were presented closely in time in one session by a single experimenter. Although
Gick and Holyoak told their subjects that the stories and problems were parts of
different experiments, there is still the possibility that the context served as a cue
whereby the stories were brought to bear on the radiation problem. That is, on
receiving the radiation problem, the subjects may have thought back to the stories
only because they had just been presented in the same session, not because they
were structurally analogous to the radiation problem. The subjects may have been
trying to figure out why they were given two tasks to do in the same session, which
may have led to the conclusion that the two tasks were related. This, in turn, could
lead them to try to relate the stories to the radiation problem. According to this
view, analogical thinking is not the cause of the facilitation of the tumor problem
brought about by prior exposure to the stories.
In a test of this hypothesis, Spencer and Weisberg (1986) replicated Gick and
Holyoak’s study, except that the context was changed from the stories to the tumor
problem. This was done by having the stories presented as part of an experiment at
the beginning of a class, conducted by a visitor unknown to the students. After the
experimenter left, the instructor presented the radiation problem as part of a class
discussion of problem solving, without making any mention of a relation between
the stories and the radiation problem. Spencer and Weisberg found that no transfer
occurred from the analogous stories to the radiation problem if subjects were not
explicitly informed of the relation between them. When the stories and critical
problem were presented in different contexts, no spontaneous noticing of their
analogous structures occurred. Catrambone and Holyoak (1985) and Keane (1985)
have recently reported results using the radiation problem that support this conclu¬
sion, as do results from studies using other target problems (Reed, Ernst, & Baner-
160 R. W. Weisberg

jii, 1974). Thus remote analogy in and of itself may not provide a basis for memory
search in uninformed subjects.
In conclusion, results from research using several different problems support the
conclusion that during problem solving only a very limited search is made of
memory. Information that can be shown to be available in memory and potentially
relevant to some problem may not be accessed if the connection of that information
to the problem is not explicit and/or direct. Human thought seems not to be of great
range and flexibility during laboratory problem solving.

Broader implications: problem solving and creativity

The research on problem solving just discussed can provide an organizational

framework for an initial examination of the literature on creativity in science and the
arts. Although the present viewpoint is nowhere near a theory of creativity, the
analysis just presented can serve to generate some expectations about what one
should expect to find in examining individual examples of creative thinking. First,
if creativity begins with a match between a new situation and one’s knowledge, then
one should see creative products beginning with rather prosaic variations on old
themes. These old themes may come from the work of others, or from the work of
the individual in question, depending on circumstances to be elaborated later.
Second, novel products evolve in small steps out of these initial attempts. No leaps
of creativity should be found for even the most creative outcomes if detailed infor¬
mation concerning the individual’s work is available. Third, this evolution does not
come about in a vacuum. That is, some prosaic initial product will not be modified
into something new unless two factors come into play: The individual must see
some inadequacy in the initial product, and the individual must know how to
overcome this inadequacy.
The examples to be considered come from science, technology, and the arts. It is
relatively obvious that a framework based on research in problem solving might
have relevance to work in science and technology, because the production of a
scientific theory or an invention usually is a “solution” to some “problem.” The
case in the arts is less clear, but in order to see how far the present point of view can
go, I shall assume that artistic creativity can also be looked on as problem solving.
This issue is discussed further by Weisberg (1986, chaps. 7 and 8).

Double helix

Watson and Crick’s double-strand helical model of the DNA molecule set off the
ongoing revolution in molecular genetics. The importance of this discovery has led
some to assume that it must have come about in a burst of creative insight, indepen¬
dent of everything that came before (Adams, 1979, pp. 60-61):
Watson and Crick relied heavily on inspiration, iteration, and visualization. Even though they
were superb biochemists, they had no precedent from which they could logically derive their
structure and therefore relied heavily on left-handed [non-logical] thinking.
Problem solving and creativity 161

However, contrary to Adams, Watson’s own autobiographical account of their

discovery (Watson, 1968) indicates that although the model was revolutionary in its
effects, it evolved in a relatively straightforward way out of the backgrounds of
Watson and Crick, in interaction with the circumstances in which they found them¬
selves (Judson, 1979). There almost certainly were precedents available, and they
used them.
Both Watson and Crick had, during their separate training, become impressed
with the importance of solving the problem of the structure of DNA. In addition,
Crick was a PhD student in the Cavendish Laboratory at Cambridge University,
which had for a long time been involved with using X-ray crystallography to
analyze the structures of molecules. Watson had seen an X-ray photograph of
crystalline DNA and had been impressed with the fact that DNA formed a crystal,
as well as with the potential of the X-ray technique to provide information concern¬
ing the structure of DNA. Thus, when Watson came to work at Cavendish in the fall
of 1951, he and Crick were kindred spirits, and they decided to try to build a model
of DNA. Data obtained from X-ray crystallography would provide one of the bases
for their work. Much important work on X-ray analysis of DNA was being carried
out by a group at King’s College, London, and Watson and Crick were familiar with
this work and in contact with those scientists before they began their own theoriz¬
ing. This, then, is one precedent.
A much more important precedent was the work of Pauling, who was world-
famous for developing models of molecular structures. Furthermore, he had re¬
cently published a model of a protein, helical in shape, that had greatly impressed
both Watson and Crick. Therefore, they decided to use Pauling’s work as the basis
for their own, and they developed a helical model of DNA. Two specific problems
arose in constructing this model: How many helical strands should it contain, and
where should the bases be located? The DNA molecule looks something like a spiral
staircase. There are two helical strands, or backbones, held together by pairs of
bases, which form the “stairs.” In Watson and Crick’s initial model, however,
there were three strands, rather than the final two, and the bases projected outward
from the backbones, rather than being between them. Watson and Crick chose a
three-strand structure for their initial model for several reasons. First, the available
X-ray evidence indicated that the molecule was thicker than a single strand. Second,
evidence reported by Franklin, from the King’s College group, also pointed to three
strands. Watson and Crick put the bases on the outside because at that time they
could not see a way to put them inside the backbones while maintaining the regular
structure shown in the X-ray pictures. Putting the bases on the outside did away
with this problem. These first steps in the discovery of the structure of DNA were
relatively straightforward: Watson and Crick took an already available model and
modified it to solve their specific problem. Nothing in the way of an intuitive leap of
creative insight is seen here.
However, Watson and Crick ran into an immediate snag when the group from
King’s College came to see their three-strand model and told them that it surely was
incorrect, based on the King’s College data concerning the amount of water DNA
162 R. W. Weisberg

contained, among other things. Thus, Watson and Crick were forced back to the
drawing board, but not all the way to the beginning. They retained the helical
structure, but changed it in response to new information that became available, in an
example of a “local search.” Two large changes took place in the structure: The
bases were moved between the backbones of the helix, and the number of strands
went from three to two.
Several pieces of evidence led to a bases-inside, backbone-outside model. First of
all, Watson and Crick could not devise a backbone-inside model that did not violate
basic laws of chemistry concerning distances between molecules. Second, Watson
saw a new X-ray photograph made by Franklin, and he could see that the backbones
were on the outside. As far as the number of strands were concerned, again there
was information that led Watson and Crick from three strands to two. First, a
number of other investigators, including Pauling himself, were working on three-
strand models, and none of them showed signs of success. Second, information in a
report by Franklin indicated that the backbone chains came in pairs and ran opposite
to each other, which led to two strands as the simplest place to begin. Thus, over
about a year, Watson and Crick moved away from their initial incorrect model to
one that turned out to be correct. This movement to the new model came about as a
result of pushes from information that made clear not only that the early model was
wrong but also in what direction the solution lay.
There was now one last element to fit in: How were the bases to be paired to
connect the backbones? As mentioned earlier, one reason that the original model
had the bases on the outside was that there seemed to be no way to fit the bases
inside without making the structure too irregular to produce the patterns shown in
the X-ray photographs. However, when Watson and Crick built their first two-
strand bases-inside model, the models of the bases were not yet available from the
machine shop, so they could begin their work with the backbones, without worrying
about the bases.
Once the two-strand backbone structure was completed, and the base models
became available, attempts were made to pair up the bases in various ways, so that
they could serve to hold the two backbones together. The initial attempts involved
pairing bases of the same type, called “like-like” pairings, but this led to pairings
of different sizes, which would produce an irregular structure, which was unaccept¬
able. According to Watson (1968, pp. 123-125), this failure using like-like pairings
led him to try various other combinations, basically by trial and error, until the
correct combinations were hit upon.
In conclusion, we have here one of the great discoveries of modem science, and
it seems to have come about in a manner very different from the romantic view
of the scientist working alone in the laboratory until a sudden insight leads to the
creative solution to the problem. In the case of Watson and Crick, their initial
work grew directly out of the work of others, and when it was shown to be
wrong, they modified it in straightforward ways, until a satisfactory solution was
arrived at.
Problem solving and creativity 163

Darwin’s theory of evolution

If one reads Darwin’s autobiography concerning the events that led to the develop¬
ment of the theory of evolution through natural selection, one gets the impression of
a sudden insight (Gruber, 1981, p. 172-173):

In October 1838, that is, fifteen months after I had begun my systematic enquiry, I happened
to read for amusement Malthus on Population, and being well prepared to appreciate the
struggle for existence which everywhere goes on from long and continued observation of the
habits of animals and plants, it at once struck me that under these circumstances favorable
variations would tend to be preserved and unfavorable ones to be destroyed. The result of this
would be the formation of a new series. Here, then, I had at last got a theory by which to
work.

Darwin’s reading of Malthus allowed him to formulate in essentially complete form

the theory of evolution through natural selection. Soon after reading Malthus,
Darwin wrote the following in his notebook (Gruber, 1981, p. 173):

Three principles will account for all 1. grandchildren like grandfathers, 2. tendency to small
change especially with physical change, and 3. great fertility in proportion to support of
parents.

On the basis of this report it would appear that Darwin had simply collected facts in
a neutral manner, without any theory, until a passage from Malthus brought every¬
thing into place. However, Darwin’s notebooks present a very different picture of
how the theory developed (Gruber, 1981). As was the case with Watson and Crick,
Darwin’s work was based on that of others (Eiseley, 1961), and the final form of the
theory only gradually evolved.
Interest in the question of how species developed obviously predated Darwin.
Indeed, the various components of Darwin’s theory had been proposed as separate
fragments by others, in some cases many years earlier (Eiseley, 1961). Religious
orthodoxy of the time assumed that all the species had been created at once, in final
perfect form, by God, and that no evolution had occurred. However, in liberal
circles, many scientists and philosophers believed that a strict interpretation of the
biblical view was incorrect and that evolution of some sort must have occurred. In
addition, there was specifically in Darwin’s family great interest in questions con¬
cerning evolution, because Darwin’s grandfather, Erasmus Darwin, had proposed a
theory of evolution based on the inheritance of acquired characteristics. Thus,
interest in the questions that occupied Darwin during his adult years probably was
stimulated early in his life.
Darwin’s higher education, at the University of Edinburgh and Cambridge Uni¬
versity, also contributed to his development. At these institutions, Darwin met
many scientists who were actively thinking about problems in evolutionary theory
and who believed in one or another variant of the then current views, such as
Lamarck’s theory of inheritance of acquired characteristics. After graduating from
Cambridge, Darwin left on a 5-year voyage aboard HMS Beagle, whose ports of
call provided him with firsthand evidence of the remarkable variety and adaptability
164 R. W. Weisberg

of species. In summary, when Darwin returned from the voyage on the Beagle, he
was a sophisticated scientist who was thoroughly familiar with current theorizing
concerning evolution and who possessed a unique collection of facts culled from his
observations during the voyage.
Contrary to the report in Darwin’s autobiography, he had developed a theory of
evolution in July 1837, 15 months before reading Malthus (Gruber, 1981). The
early theory is based on the concept of the “monad,” an idea that was not original
to Darwin. Monads were believed to be simple living particles that were constantly
springing spontaneously into life: Each monad develops into a whole group of
related species, and when the monad dies, all its species die at once. The monad’s
developing species respond to environmental forces with adaptive changes, in a
variation on the theorizing of Darwin’s grandfather. Thus, contrary to Darwin’s
claims, he had a theory long before reading Malthus, and this early theory was
firmly rooted in the scientific theorizing of the time. As with Watson and Crick,
Darwin’s theorizing had precedents, and he used them relatively directly. Only
gradually over the next 15 months did Darwin develop the theory of evolution based
on natural selection. This theory evolved out of the monad theory, as it gradually
became apparent to Darwin that the monad theory would not do the job of explain¬
ing evolution of species.
As one example of the ways in which Darwin’s thinking changed, in the monad
theory it was assumed that variation among members of a species was in response to
environmental change. In the theory of natural selection, variation is assumed to
occur independent of the environment, with some of these variations being better
adapted than others, and so on. This shift in Darwin’s view may have been brought
about in part by his analysis of the results he observed during the voyage of the
Beagle, which showed that great variation came about in nature independent of
environmental change. Therefore, rather than being a result, variation became a
given.
Thus, when Darwin read Malthus, his ideas had been changing for over a year, to
the point that Malthus could make a crucial impression on him. Malthus triggered
no great leap of insight, but rather provided the capstone on a reasoning process that
had moved in small steps from a theory that was a rather simple variation on already
known themes to a theory that would change the world.

Edison’s invention of the kinetoscope

Edison’s technological creativity produced thousands of inventions, two of the most

important of which were a device for storing and reproducing sound (a phonograph)
and a device for storing and reproducing moving visual information (the ki¬
netoscope - the first moving pictures). The documentation surrounding these two
inventions, including sketches in Edison’s notebooks as well as his early patent
applications, indicates that Edison’s thought processes also may have followed the
pattern outlined earlier (Jenkins & Jeffrey, 1984).
Problem solving and creativity 165

The phonograph was invented in Edison’s laboratory in 1877, and its demonstra¬
tion in the next year made Edison world-famous. One preliminary sketch of the
phonograph, made on November 29, 1877, shows a cylinder with its long axis
horizontal as the central piece of the apparatus. The sound information was carried
in a spiral groove impressed into the surface of the cylinder, and a moving stylus
was used to retrieve that information from the rotating cylinder. A second sketch,
made a few days later, indicates that Edison and his associates had been considering
at least three possible mechanisms for storing sound information: a cylinder, a flat
disk, and a narrow tape. The final phonograph was one of these possibilities.
About 10 years later, Edison produced the kinetoscope, a machine that enabled a
viewer to see moving images. The kinetoscope served a purpose very different from
that of the phonograph, and the kinetoscope was very different in design and
appearance from the earlier invention. The kinetoscope consisted of an upright
cabinet with a viewing peephole at the top. The viewer looked down into the cabinet
and saw the moving image. Film was wound on many rollers inside the cabinet,
allowing it to be drawn under the peephole, to provide the moving image for the
viewer.
Although on the surface the phonograph and the kinetoscope appear to be entirely
different devices, analysis of Edison’s preliminary work on the kinetoscope indi¬
cates that in reality the former served as the basis for the initial development of the
latter. Edison’s earliest preliminary patent application (called a “caveat”), made in
October 1888 for the kinetoscope, contained a sketch that looked nothing like the
final version. Rather, this sketch contained an apparatus that had as its central piece
a single cylinder that rotated around its horizontally oriented long axis. On the
cylinder was a spiral of images that were viewed through a moving eyepiece as the
cylinder rotated. Thus, the earliest version of the kinetoscope seems to have been
based directly on the phonograph. The written text of the caveat, of which the
sketch described was a part, makes clear the relationship between the two in¬
ventions (Jenkins & Jeffrey, 1984, p. 5):

I am experimenting upon an instrument which does for the Eye what the phonograph does for
the ear, Which is the recording and reproduction of things in motion. . . . The invention
consists in photographing continuously a series of pictures occuring [sic] at intervals which
intervals are greater that [sic] eight per second, and photographing these series of pictures in a
continuous spiral on a cylinder or plate in the same manner in which sound is recorded on the
phonograph.

In sum, there is little doubt that the kinetoscope evolved directly out of the
phonograph, as Edison took the idea of the horizontal rotating cylinder and tried to
use it in a different domain. Although the finished products were entirely different
in form, the documents provide evidence concerning the intermediate steps that led
from one to the other. Furthermore, there is evidence that Edison was forced to
abandon his early phonograph-based kinetoscope because of problems in fashioning
the needed spiral of images, which led to the use of film on rolls.
In a historical analysis of Edison’s notebooks, including many drawings, Jenkins
166 R. W. Weisberg

(1983) has noted that Edison used a small number of ideas or forms in attacking
many problems. Perhaps the most frequent example is the use of the cylinder, as we
have just seen, and other such forms were the tuning fork and curved ratchet pawl.
Jenkins traces the importance of the cylinder in Edison’s inventions to his early
work experience as an operator in telegraph offices, which contained mechanical
devices using cylinders. In addition, Edison’s work on newspapers exposed him to
printing presses using rotating cylinders. Finally, his work in machine shops gave
him much experience with the lathe, which involves a rotating piece of material
being cut into cylindrical form by a cutting tool directly analogous to the stylus in
Edison’s inventions.
In conclusion, this analysis indicates that Edison’s approach to the problem of
reproducing moving visual images was based on his earlier work in a related area,
which in turn was based on his earlier work experience. Furthermore, the changes in
direction that led him away from the initial conception of the kinetoscope seem to
have been responses to problems that developed as he attempted to bring his early
conception to reality.
The pattern shown in Edison’s work on the kinetoscope is somewhat different
from that seen in the discoveries of Watson and Crick and of Darwin. In these latter
cases, the initial conception was based relatively directly on the work of others. In
the case of Watson and Crick, Pauling’s work with helical models served directly as
the basis for their work, and Darwin’s monad theory was also rather directly based
on ideas of others. In the case of Edison, on the other hand, the kinetoscope was
based on his own earlier work on the phonograph. This earlier work can in turn be
traced back to Edison’s earlier experiences with cylinders employed in designs by
others, but by the time Edison started to work on the kinetoscope he already had
available a stock of his own ideas, which he then used as the basis for attacking new
problems. However, although Edison had another layer of experience to work with,
his own extensive experience, the same basic principles are exemplified in all the
work discussed so far. Another example of the development and use of one’s own
work as the basis for dealing with a new problem can be seen in Picasso’s painting
of his great mural Guernica.

Picasso’s Guernica

In 1936, during the Spanish civil war, Picasso was commissioned by the Spanish
govemment-in-exile to produce a work for the government’s pavilion in an interna¬
tional exposition to take place in Paris in June 1937. The bombing by the Nazis of
the Basque town of Guernica on April 27, 1937, stimulated Picasso to fulfill this
commission by producing one of his most famous works, the large mural Guernica.
This work is particularly important to students of artistic creativity, because Picasso
catalogued all of his preliminary work for Guernica, some 40 pieces, and he also
had the mural photographed four times as he painted it. If one examines this
Problem solving and creativity 167

information concerning Guernica as work-in-process, one sees the gradual develop¬

ment of the final form of the work (Amheim, 1962; Blunt, 1969).
Guernica is painted on a large scale - it measures over 25 feet by 11 feet high.
There is no color; the painting is done in shades of gray, ranging from white to
almost black. The central figure is a horse with its head raised in a scream of agony.
Above the horse’s head is a sun with a light bulb superimposed on it. A bird flies up
toward this sun. On the left is a large bull, whose body faces the middle of the
scene, but whose head is turned away. Beneath the head of the bull is a mother, her
head tilted back and her mouth open in a scream, holding a dead baby. To the right
of center, a woman leans out of a burning building, holding a lamp. On the upper
far right, a woman with burning clothes is falling out of a building, while below a
woman rushes into the scene. Along the bottom left are the head and arms of a
statue of a warrior, in whose hand is a broken sword.
Picasso began to work on sketches for the mural almost immediately after receiv¬
ing news of the bombing. The first few sketches contain some of the main charac¬
ters in the final mural (the bull, horse, and woman holding a light), but their
organization changes greatly from one sketch to another, and some of them (e.g.,
the woman holding the light) are not present in all the sketches. In addition, the
main characters, as well as several of the other characters, are subjects of separate
paintings and drawings, in what seem to be Picasso’s attempts to determine exactly
how each character should look. In addition, there are elements that are introduced
in the middle of the work, but are then removed before completion. One gets the
feeling from this work that Picasso had some idea of what he wanted to include in
the work, but the specifics were not clear to him when he started. For example,
when he began to put paint to canvas, the center of the mural was dominated by a
defiant upraised arm, with clenched fist, that was not present in the preliminary
sketches. However, Picasso gradually modified it until it disappeared from the final
version, as he kept revising things. Indeed, Picasso was making revisions even after
the mural was hung in the Spanish pavilion in the Paris exposition, and Blunt
believes that one of these revisions was stimulated by another work on display in
that pavilion.
It is also of note that the characters and their organization in Guernica are closely
related to much other work done by Picasso at about the same time, as well as
earlier. The horse and bull, for example, appear again and again in works on the
bullfight, among others. A girl or woman holding a light, while confronting a bull
or minotaur, is also present in many of Picasso’s works. Also, an etching, Mino-
tauromachy, produced a few years before Guernica, contains essentially all the
characters that later appeared in Guernica, and in a very similar organization. Thus,
Picasso developed Guernica out of themes that had appeared in his earlier works.
Furthermore, Blunt has traced some of the specific gestures and configurations used
in Guernica to specific sources in the works of earlier artists with which Picasso was
familiar.
168 R. W. Weisberg

In conclusion, Picasso’s production of Guernica parallels closely Edison’s pro¬

duction of the kinetoscope. In both cases the initial conception of the work came
from earlier work of the individual, and the initial conception evolved into the final
version. In addition, the early work of the individual can be seen to have sources in
the work of others.

Conclusions

These examples, though highly selective, are still of great importance for theories of
creativity. They are cases of creativity of the highest sort, and they are also cases in
which some documentary evidence exists concerning the thought processes in¬
volved. In all these cases, and in others not discussed because of lack of space
(Weisberg, 1986, chaps. 6 and 7), one sees nothing like the creative leaps postu¬
lated by the genius view. Rather, in all these cases, some new product - a painting,
a scientific theory, or an invention - began as a rather straightforward extension of
earlier work. The initial conception then underwent modification, until something
new emerged. The important point is that in these cases there is some sort of
evidence available that goes beyond the creator’s statements concerning the pro¬
cesses involved, and in each case nothing particularly extraordinary occurred from a
psychological point of view. This is not to say that these cases do not involve
extraordinary occurrences from a scientific, technological, or artistic point of view,
but that the cognitive processes involved were not extraordinary.
Sternberg and Davidson (1982, 1983; Davidson & Sternberg, 1984) have re¬
cently argued that intellectual giftedness may depend in part on a special set of skills
that constitute what they call insightful thinking. According to their view, major
intellectual advances, in the sciences and the arts, involve insights, and the ability to
carry out insightful thinking may be independent of IQ. Sternberg and Davidson
have examined this ability by having subjects attempt to solve problems of various
sorts, and then examining the relationships between performance on these problems
and other measures of intellectual functioning. The components of this hypoth¬
esized insight ability are selective encoding, selective combination, and selective
comparison.
Selective encoding involves the ability to pick out the relevant information from a
problem and to ignore the irrelevant. According to Sternberg and Davidson, most
interesting problems involve large amounts of information, most of which is irrele¬
vant to the solution. Insight involves the ability to determine just which information
will be useful in the long run. Fleming’s discovery of penicillin is presented as an
example of selective encoding, because it depended on Fleming’s selectively encod¬
ing the information that bacteria had been destroyed by a mold that had accidentally
come into contact with the dish in which the bacteria were growing. Rather than
encoding the result as a failed experiment, Fleming encoded it as destruction of
bacteria, which resulted in a new and valuable way of conceiving of the result.
Selective combination involves putting relevant pieces of information together in
Problem solving and creativity 169

just the right way. That is, once the relevant information has been selectively
encoded in some problem, it is still necessary to develop the solution out of this
information. Sternberg and Davidson cite Darwin’s formulation of the theory of
evolution as an example of selective combination. According to Sternberg and
Davidson, Darwin had available all the facts that ultimately were to form the theory.
What was needed, according to Sternberg and Davidson, was a coherent way of
putting the facts together.
Finally, selective comparison involves relating some new information to one’s
preexisting knowledge in just the right way. One uses one’s knowledge to under¬
stand the new information, and productive use of one’s knowledge depends on
realizing just how the new information can be matched with what one already
knows. As an example of selective comparison, Sternberg and Davidson cite Kek¬
ule’s discovery of the benzene ring through the use of the image of the snake biting
its tail. Presumably, if Kekule had not used just that metaphor, he would not have
had his insight.
Sternberg and Davidson base their discussion of insight at least in part on their
analysis of several great discoveries, discoveries that presumably were based on
insight. From my perspective, the situation is much more complicated than Stern¬
berg and Davidson acknowledge. As one example, it does not seem to be the case
that Darwin was in possession of all the relevant information years before he
produced the theory of natural selection. Gruber (1981) presents a strong case that
before he could produce the final theory, Darwin’s thinking had to change in crucial
ways. These changes were such that it does not seem to be true that everything was
in place and simply waiting for a framework to come along. As will be seen shortly,
Kekule’s discovery of the benzene ring also may have been more complicated than
Sternberg and Davidson’s analysis would lead one to believe.
In sum, although Sternberg and Davidson’s research may have isolated a mode of
thinking that could be called insight, it is not clear to me that this sort of thinking is
involved in creative discovery of the sort discussed in this chapter.

Experience and creativity: influences and practice

The present point of view assumes that creative products begin with an individual
using personal experience as the basis for approaching some problem. Past experi¬
ence often is seen in the form of “influences,” in which one can see traces of the
work of some earlier individual in the work of a creative artist or scientist. Exam¬
ples of this have already been seen in the case studies just discussed. Watson and
Crick were influenced by Pauling, and Darwin was influenced by several of his
teachers, among others. Needless to say, case studies of great musicians, artists,
and scientists are filled with examples of even the greatest individuals being influ¬
enced by the work of others. Specific compositions of Bach, Beethoven, Haydn,
and Mozart can, for example, be traced to earlier works by others (Weisberg, 1986,
chaps. 6 and 7). Another sort of influence can be seen especially in literature and
170 R. W. Weisberg

drama, where specific life experiences serve as the basis for the artist’s work.
Again, examples abound, ranging from Dostoyevsky’s use of his prison experience
in Crime and Punishment to F. Scott Fitzgerald’s use of his life experiences in
Tender Is the Night. (Weisberg, 1986. chap. 7).
The development that leads from dependence on the work of others to the produc¬
tion of original work of one’s own may extend over a significant period of time in
the life of the creative individual. A recent study of musical composition by Hayes
(1981, chap. 10) estimated that it takes approximately 10 years of composing
experience (10 years of practice) before one will produce a masterwork. The early
compositions of even the greatest composers often are of value only historically,
because musically they are derivative of the work of others. Even Mozart, who
began his musical studies around the age of four, did not, according to Hayes’s
criteria, produce a masterwork until he had been working for 12 years. Thus, an
evolution away from the prosaic seems to occur even in the work of the greats.

Creativity and myth

This chapter began with three self-reports from individuals of acknowledged

creativity: Coleridge, Mozart, and Kekule. These are but a small sample of many
such reports, all of which are remarkably similar in content (Ghiselin, 1952). These
reports obviously contradict the present evolutionary view of creative thinking, and
therefore a brief discussion of their validity will be of value. It should first be noted
that those three self-reports are contradicted by the case studies discussed in the last
section. In the case studies, in which detailed information was available, it was
found that the thought processes involved were not of the sort postulated by the
genius view. No great leaps of thought were found to occur. As mentioned earlier,
there are other case studies available that also do not support the genius view. From
this perspective, then, uncorroborated self-reports must be brought into question.
Even greater damage to the genius view comes about when these self-reports are
examined in detail, because questions can be raised about the validity of all three of
them.

Coleridge and “Kubla Khan”

Coleridge claimed that “Kubla Khan” came to him complete, with no effort on his
part, in an opium-induced dream or stupor. He was able to commit only a fragment
of the poem to paper before he was interrupted, which caused him to lose the rest of
the poem forever, or so he claimed. There is other evidence that leads one to believe
that almost nothing Coleridge reported about the genesis of “Kubla Khan” is true
(Schneider, 1953). Most important, another version of the poem exists, and evi¬
dence indicates that this other version is earlier than the version actually published.
This, of course, leads to the conclusion that Coleridge worked on the poem before
publishing it, which belies his claim that he simply set down the lines that spon-
Problem solving and creativity 171

taneously came to him in his stupor. Schneider also argues that opium would not
have induced the sort of visions that Coleridge claimed to have experienced. In
addition, Coleridge was notorious for not telling the truth concerning his work.
Finally, I find it difficult to believe that a poet in the throes of writing down a
miraculously given poem would allow himself to be interrupted for an hour by a
visitor on some “business” errand. Surely Coleridge would have asked the man to
wait until he was finished transcribing the two to three hundred lines that had come
to him.
Schneider has proposed that Coleridge fabricated the account of the composition
of “Kubla Khan” in order to make what was in reality only a fragment of a poem
more interesting to his public. If Coleridge could not complete the poem, argues
Schneider, then transforming the essentially worthless fragment into all that re¬
mained of an interrupted burst of inspiration might make it something valuable.
Whether or not Schneider’s speculations are correct, there is no doubt that Cole¬
ridge’s report is of little value to students of creative thinking.

Mozart’s letter

The excerpt from Mozart’s letter is similar in content to Coleridge’s report, in that
Mozart claimed that his compositions grew into completed works without his con¬
scious participation. He then wrote down the complete works. Again there is evi¬
dence that this report should not be believed. First, there is strong evidence that
“Mozart’s letter” was not written by Mozart (Deutsch, 1964). This conclusion,
though long accepted by musicologists (Anderson, 1966), has not penetrated into
the psychological literature. Also, and more important, recent analyses of Mozart’s
original manuscripts do not indicate that he simply committed to paper already
complete works (Sloboda, 1984, pp. 112-114). So here again the validity of the
self-report is called into question.

Kekule’s “dream”

The final self-report was that of Kekule, concerning an alleged dream in which he
hit upon the circular structure of benzene through the medium of the image of a
snake biting its own tail. In this case also, questions can be raised about the ordinary
interpretation of this report. First, there is a question concerning whether Kekule
was even dreaming (Rothenberg, 1979, pp. 395-396). The original report is in
German, and Kekule used the term “habschlaft” (half-sleep) to describe his state,
which may mean more that he was daydreaming, perhaps sitting transfixed in front
of the fire, rather than truly dreaming. There is also a question concerning whether
or not Kekule was imagining snakes during this incident. In his report, he says that
he saw the strings of atoms in “snakelike motion.” The use of the term “snake¬
like” indicates that the strings of atoms had not become transformed into snakes,
because he would then simply have referred to snakes. His use of “snakelike”
172 R. W. Weisberg

indicates that he was distinguishing between the strings of atoms and snakes. If this
analysis is correct, then Kekule’s saying that one snake had seized hold of its tail
must only be figurative. Because he had called the motion of the strings snakelike,
then when one of the strings formed into a ring, he described it as biting its own tail.
Thus, although Kekule may have used visual imagination to think about the
structure of benzene, he may not have been imagining snakes, and he may not have
been dreaming. This analysis is relevant to Sternberg and Davidson’s (1983) discus¬
sion of Kekule, cited earlier.

Conclusions

Although this critical review of the three self-reports is no less selective than the
earlier analysis of case studies, once again this does not diminish the main point.
The self-report examples chosen are among the most often cited in support of the
genius view of creativity. The fact that the validity of these examples can be called
into question, combined with the earlier analysis of better-documented case studies,
indicates that the genius view of creativity may be based on myth rather than fact.

Can one’s creativity be increased?

A helpful way to provide a summary of the main points raised in this chapter is to
consider the problem of increasing an individual’s creativity. There is a great
industry built on the premise that many people need help in order to think cre¬
atively. Many books provide hints as to how to increase creativity (Adams, 1979; de
Bono, 1967), and many methods are taught in courses and seminars that are at¬
tended by thousands of individuals each year; see Nickerson, Perkins, and Smith
(1985) for a review. The discussion in the present chapter, on the other hand, has
led to the conclusion that the thought processes involved in great creative acts are no
different than those involved in the things we all do every day. This means that
creative thinking must be omnipresent in all of us, which means that it is neither
necessary nor possible to increase anyone’s capacity to be creative. Therefore, in
order to consider the question whether or not one’s creativity can be increased, one
must reformulate the question, in order to put it into a form that can lead to an
answer.
Why would one approach an expert in creative thinking in order to have one’s
creative capacity increased? One reasonable possibility is that one has had problems
dealing with some aspect of one’s life, an aspect that presumably could benefit from
“creative thinking.” I might be concerned that students find my lectures boring,
and so I wish to become more creative in thinking of ways to make my lectures
interesting. A salesperson might be interested in thinking of new ways to convince
prospective buyers to purchase some product. A scientist might be concerned be¬
cause of having made little progress in solving some important scientific problem. If
Problem solving and creativity 173

we now focus on these sorts of specific questions, there may be some information
from the earlier discussion that might be helpful.
Great innovations in science and the arts were almost invariably produced by
individuals who possessed strong motivation and persistence. Such individuals as
Darwin, Picasso, Edison, Einstein, and Beethoven, among many others, spent
lifetimes working in their chosen fields, and their genius came about at least in part
as a result of this lifetime of work. Commitment provides sufficient time for the
small changes that occur as one gathers experience in some domain to evolve into
something truly original and innovative. Commitment also brings with it a tendency
to spend much time thinking about one’s work, which also raises the possibility that
something novel will develop. Time spent thinking about one’s work can also
increase the likelihood that a potentially relevant external stimulus will be noticed
(Weisberg, 1986, chap. 1).
A second important conclusion in the present chapter is that creative products are
firmly based on what came before. True originality evolves as the individual goes
beyond what others had done before. This might mean, perhaps paradoxically, that
in order to produce something new, one should first become as knowledgeable as
possible about the old. This serves to provide the background so that the individual
can begin to work in an area and also serves to provide ways in which to modify
early products that are not satisfactory.
These two aspects of creative work, commitment and expertise within one’s own
area, are neither profound nor novel. All scientists and artists have extensive train¬
ing, either formally or informally, and very few individuals make a mark in the
world without a relatively long commitment to an area beyond their actual training.
It might be hoped that something more exciting could be presented concerning the
teaching of creative thinking, but the issues are too complex to permit straightfor¬
ward answers. Consider one personality characteristic that is sometimes cited as
inhibiting creativity: fear of taking risks. In a popular book that surveys various
ways of increasing creativity, Adams (1979, chap. 3) discusses fear of taking risks
as a factor that may interfere with creativity. An individual may come up with a
novel and possibly adequate solution to a problem, but fear of being incorrect stops
the individual from making it public. Thus, one way to increase one’s creative
capacity would be to increase one’s capacity for taking risks. Adams attempts to do
this by trying to convince the reader that nothing really awful will occur if one is
wrong, and so taking a risk and failing is better than doing nothing.
Reconsideration of some of the case studies from this chapter, however, indicates
that risk taking may not be related in a simple way to producing a creative product.
Watson and Crick seemed to take a risk, choosing a helical structure for their
model, and it paid off in their being first to develop a model of DNA. On the other
hand, such a way of approaching a problem probably would not have helped Darwin
develop his theory. In the case of the theory of evolution, the final theoretical
structure seems to have required a number of components coming together, and all
 174 R. W. Weisberg

of them were needed before anything coherent could emerge. Therefore, one cannot
simply say: Take more risks. Whether or not one should take a risk may depend on
the problem on which one is working. There may be as many ways to be creative as
there are problems, in the broadest sense, that humans can ponder. If that is so, then
there may be no simple way to change people to make them more creative.
In summary, this chapter has attempted to take seriously several implications
arising from laboratory research in problem solving, to provide a framework for
analyzing creative thinking in science, technology, and the arts. Laboratory re¬
search in problem solving indicates that, contrary to the genius view, creative
thinking may require neither extraordinary individuals nor extraordinary thought
processes. Analysis of several case studies also shows that, perhaps surprisingly,
several great creative achievements also seemed to involve rather ordinary thought
processes. This convergence of the results from two such disparate domains indi¬
cates that the genius view of creativity, although it has broad and deep roots in our
culture, may be basically misdirected.

References
y/ Adams, J. L. (1979). Conceptual blockbusting (2nd ed.). New York: Norton.
Anderson, E. (1966). The letters of Mozart and his family (2nd ed.). New York: St. Martin’s Press.
Amheim, R. (1962). Picasso’s Guernica: The genesis of a painting. Berkeley: University of California
Press.
Blunt, A. (1969). Picasso’s “Guernica.” New York: Oxford University Press.
Bourne, L. E. Ekstrand, B. R., & Dominowski, R. L. (1971). The psychology of thinking. Englewood
Cliffs, NJ: Prentice-Hall.
Bransford, J. (1979). Human cognition. Learning, understanding and remembering. Belmont, CA:
Wadsworth.
Burnham, C. A., & Davis, K. G. (1969). The 9-dot problem: Beyond perceptual organization. Psycho-
nomic Science, 17, 321-323.
Catrambone, R., & Holyoak, K. (1985, August). The role of schemas in analogical problem solving.
Paper presented at American Psychological Association convention, Los Angeles.
Coleridge, S. T. (1816). Prefatory note to Kubla Khan. In B. Ghiselin (Ed.), The creative process (pp.
84-85). New York: Mentor.
Davidson, J. E., & Sternberg, R. J. (1984). The role of insight in intellectual giftedness. Gifted Child
Quarterly, 28, 58-64.
de Bono, E. (1967). New think: The use of lateral thinking in the generation of new ideas. New York:
Basic Books.
DeGroot, A. (1966). Perception and memory versus thought: Some old ideas and recent findings. In B.
Kleinmuntz (Ed.), Problem solving: Research, method, and theory (pp. 19-50). New York: Wiley.
Deutsch, O. E. (1964). Spurious Mozart letters. Music Review, 25, 120-123.
Duncker, K. (1945). On problem-solving. Psychological Monographs, 58, No. 5, Whole No. 270.
Eiseley, L. (1961). Darwin’s century. Evolution and the men who discovered it. New York: Anchor
Books.
Ghiselin, B. (Ed.). (1952). The creative process. New York: Mentor.
Gick, M. L., & Holyoak, K. (1983). Schema induction and analogical transfer. Cognitive Psychology,
15, 1-38.
Gilhooly, K. J. (1982). Thinking. Directed, undirected and creative. London: Academic Press.
Greeno, J. G. (1980). Trends in the theory of knowledge for problem solving. In D. T. Tuma & R. Reif
Problem solving and creativity 175

(Eds.), Problem solving and education: Issues in teaching and learning. Hillsdale, NJ: Lawrence
Erlbaum.
Gruber, H. (1981). Darwin on man (2nd ed.). University of Chicago Press.
Guilford, J. P. (1950). Creativity. American Psychologist, 5, 444-454.
Guilford, J. P. (1959). Traits of creativity. In H. H. Anderson (Ed.), Creativity and its cultivation (pp.
142-161). New York: Harper.
Hayes, J. R. (1981). The complete problem solver. Philadelphia: Franklin Institute Press.
Jenkins, R. V. (1983). Elements of style: Continuities in Edison’s thinking. Annals of the New York
Academy of Sciences, 424, 149-162.
Jenkins. R. V., & Jeffrey, T. E. (1984). Worth a thousand words: Nonverbal documents in editing.
Documentary Editing, 6, 1-8.
Judson, H. F. (1979). The eighth day of creation: Makers of the revolution in biology. New York: Simon
& Schuster.
Keane, M. (1985). On drawing analogies when solving problems: A theory and test of solution of
generation in an analogical problem-solving task. British Journal of Psychology, 76, 449-458.
Koestler, A. (1964). The act of creation. New York: Macmillan.
Kohler, W. (1976). The mentality of apes. New York: Liveright.
Luchins, A. S., & Luchins, E. H. (1959). Rigidity of behavior. Eugene: University of Oregon Press. <£=■
Lung, C.-T., & Dominowski, R. L. (1985). Effects of strategy instructions and practice on nine-dot
problem solving. Journal of Experimental Psychology: Learning, Memory, and Cognition, 11,
804-811.
Martindale, C. (1981). Cognition and consciousness. Homewood, IL: Dorsey.
Metcalfe, J. (1986). Feeling of knowing in memory and problem solving. Journal of Experimental
Psychology: Learning, Memory, and Cognition, 12, 288-294.
Mozart, J. C. W. A. (1964). A letter. In B. Ghiselin (Ed.), The creative process (pp. 44-45). New York:
Mentor.
Newell, A., & Simon, H. (1972). Human problem solving. Englewood Cliffs, NJ: Prentice-Hall.
Nickerson, R. S., Perkins, D. N., & Smith, E. E. (1985). The teaching of thinking. Hillsdale, NJ:
Lawrence Erlbaum.
Osborn, A. (1953). Applied imagination (rev. ed.). New York: Charles Scribner’s Sons.
Perfetto, G. A., Bransford, J. D., & Franks, J. J. (1983). Constraints on access in a problem solving
context. Memory and Cognition, 11, 24-31.
Perkins, D. (1981). The mind’s best work. Cambridge, MA: Harvard University Press.
Reed, S. K., Ernst, G. W., & Banerjii, R. (1974). The role of transfer between similar problem states.
Cognitive Psychology, 6, 436-450.
Rothenberg, A. (1979). The emerging goddess. University of Chicago Press.
Scheerer, M. (1963). Problem-solving. Scientific American, 208, 118-128.
Schneider, E. (1953). Coleridge, opium, and Kubla Khan. University of Chicago Press.
Simon, H. (Ed.) (1979). Models of thought. New Haven, CT: Yale University Press.
Sloboda, J. (1984). The musical mind. New York: Oxford University Press.
Spencer, R. M., & Weisberg, R. W. (1986). Context-dependent effects on analogical transfer. Memory
and Cognition, 14, 442-449.
Sternberg, R. J., & Davidson, J. E. (1982). The mind of the puzzler. Psychology Today, 16, 37-44.
Sternberg, R. J., & Davidson. J. E. (1983). Insight in the gifted. Educational Psychologist, 18, 51-57.
Watson, J. (1968). The double helix. New York: Signet.
Weisberg, R. W. (1980). Memory, thought, and behavior. New York: Oxford University Press.
Weisberg, R. W. (1986). Creativity. Genius and other myths. New York: Freeman.
Weisberg, R. W., & Alba, J. W. (1981). An examination of the alleged role of “fixation” in the
solution of several “insight” problems. Journal of Experimental Psychology: General, 110, 169—
192.
Weisberg, R. W., & DiCamillo, M. A. (1986). Multiple memory searches as the basis for restructuring
in an insight problem. Unpublished paper. Temple University.
176 R. W. Weisberg

Weisberg, R., DiCamillo, M., & Phillips, D. (1978). Transferring old associations to new problems: A
nonautomatic process. Journal of Verbal Learning and Verbal Behavior, 17, 219-228.
Weisberg, R. W., & Suls, J. (1973). An information-processing model of Duncker’s candle problem.
Cognitive Psychology, 4, 255-276.
Wertheimer, M. (1959). Productive thinking. New York: Harper & Row.
7 A computational model of scientific insight

Pat Langley and Randolph Jones

Creativity lies at the heart of the scientific process. Although much of science
involves the dreary application of well-worn methods, true progress requires an act
of discovery. In some cases, these discoveries take the form of insight, in which
previously unseen and unexpected connections suddenly reveal themselves to the
mind.
Introspectively, the moment of insight often has a “mystical” quality, and this
has led many to assume that the process lies outside the realm of human understand¬
ing. Early theories of scientific insight shared in this opinion, relying heavily on
notions of unconscious (and thus noninspectable) processing. But in the past few
dacades, cognitive psychology and artificial intelligence have made significant
strides in understanding the nature of human cognition. It seems only natural to
apply their methods to develop a “process” explanation of this intriguing pheno¬
menon.
In this chapter we present just such a computational theory of scientific insight.
We begin by recounting some well-known examples of the process, along with
some early theories that attempted to account for the phenomenon. We also review
some more recent attempts to explain insight in process terms, but our reservations
about these models have led us to develop an alternative theory. Our framework
builds on two separate lines of research in cognitive science: on reasoning by
analogy and on qualitative mental models. Thus, we also review some work in these
areas before moving on to the details of our model.

The advantages of cognitive simulation

Before addressing the substantive issues, we should briefly consider our meth¬
odological assumptions. One of our basic tenets is that the construction of cognitive
simulations can improve our understanding of human behavior. A cognitive simula¬
tion is simply a computer program that is intended to model human cognitive
processes in some area. This approach has proved successful in a wide variety of

This research was supported by Contract N00014-84-K-0345 from the Information Sciences
Division, Office of Naval Research. We would like to thank Bemd Nordhausen, Donald
Rose, Rogers Hall, and Jaime Carbonell for discussions that led to the ideas in this chapter.

177
178 P. Langley and R. Jones

domains, including problem solving (Newell & Simon, 1972), vision (Marr, 1982),
natural language (Schank & Abelson, 1977), and memory (Anderson & Bower,
1973).
The cognitive simulation approach has a number of advantages over more tradi¬
tional psychological methods. First, the act of constructing a running computer
program ensures that one’s theory is internally consistent. Second, one can deter¬
mine the consequences of changing a theory by adjusting the computational model
and observing the new behavior. Most important, it forces one to think in terms of
specific representations of knowledge and to specify explicitly processes for manip¬
ulating those representations. This leads to more specific - and thus more testable -
models of cognitive behavior. We refer the reader to Newell and Simon (1972) and
Anderson (1976) for additional discussion of this methodology.
The goal of our research is to construct a running cognitive simulation of scien¬
tific insight. Although we have not yet achieved that goal, we believe the very act of
thinking in process terms has revealed aspects of insight that we otherwise would
have missed.

The problem space hypothesis

Much of the research within the cognitive simulation approach has relied on what
Newell (1980) has called the problem space hypothesis. This states that all cognitive
behavior involves search through some problem space. A problem space is com¬
posed of a set of problem states, including the initial state from which search
begins. New states are generated by applying operators to existing states, letting
one systematically explore the space until the goal state has been reached.
As an example, suppose we wanted to solve some problem in linear algebra. The
initial state might be a set of n equations in n unknowns, such as

2x + 3y = 8
3x — 6y — —9

In this case, our goal would be to find some value for each unknown. There are two
operators for generating new states - adding two equations together and multiplying
an equation by a constant. An intermediate state for the foregoing problem might
include the equations

4x + 6y = 16
3x — 6y = —9

By applying the right operators in the right order, we would eventually reach the
goal state, which would tell us that x = 1 and y — 2.
Unfortunately, the problem spaces for most interesting tasks are combinatorial in
nature, so that many alternative paths present themselves. One response is to carry
out an exhaustive search of the problem space, but this rapidly becomes un¬
manageable for even simple domains. A more reasonable approach is to carry out a
Computational model of scientific insight 179

heuristic search of the problem space, using rules of thumb to suggest likely states
to expand and likely operators to select. This approach is not guaranteed to find an
optimal solution, but it is likely to produce an acceptable solution in reasonable
time. Human problem solvers appear to rely heavily on heuristic search methods.
The problem space hypothesis has been quite successful in studies of artificial
intelligence and cognitive science, and we shall see later that most explanations of
insight have been formulated within this framework. In fact, the problem space
approach has become so popular in some circles that many view it as “truth” rather
than as a hypothesis. Nevertheless, one can imagine competing frameworks for
describing cognition, and, as we shall see, our theory of scientific insight incorpo¬
rates such an alternative approach, based on the joint notions of mental models and
reasoning by analogy.

The phenomenon of scientific insight

The popular view of science assumes that progress occurs through methodical
collection of data and careful inferences from those observations. Although certain
scientific work occurs in this mode, real progress often seems to require a “leap of
intuition” or a “flash of insight” in which an old problem is suddenly seen in a
different light. Let us consider some examples of this phenomenon.
Probably the most famous instance of scientific insight is Archimedes’ discovery
of the principle of displacement (Dreistadt, 1968). The Greek scholar had been
given the problem of determining if the king’s crown was pure gold or if the gold
was mixed with silver. Knowing the density of gold and the weight of the crown, he
needed only to find its volume in order to check for purity. But the crown’s shape
was irregular, and he could not measure its volume without melting it down.
Archimedes worked on the problem for some time without finding a solution. Then,
as he lowered himself into a bath, he noticed that the water level rose simul¬
taneously. With this came the realization that any object displaces its own volume
when submerged in a liquid and that this provided the means for measuring irregular
volumes.1
Another well-known example of scientific insight is Louis Kekule’s discovery of
the ring structure of the benzene molecule. Kekule tried for some time to identify a
structural model that would account for benzene’s chemical makeup. Finally, he sat
down by the fire and began to doze (Dreistadt, 1968; Farber, 1966). In his sleepy
state he watched the smoke rising from the fire, “twisting in a snakelike motion.”
At this point, one of the snakes took its own tail in its mouth, creating a ring. In a
sudden flash, Kekule realized the molecule must be structured as a ring.
Insights seem to be fairly common in mathematics, and the eminent French
mathematician Henri Poincare (1952) reported a number of his own insights in a
lecture at the Societe de Psychologie in Paris. In one particularly striking example
he detailed his discovery of an expression for Fuchsian functions (Poincare, 1952,
p. 53):
180 P. Langley and R. Jones

At this moment I left Caen, where I was then living, to take part in a geological conference
arranged by the School of Mines. The incidents of the journey made me forget my mathe¬
matical work. When we arrived at Coutances, we got into a [bus] to go for a drive, and, just
as I put my foot on the step, the idea came to me, though nothing in my former thoughts
seemed to have prepared me for it, that the transformations I had used to define Fuchsian
functions were identical with those of non-Euclidean geometry.

This case differs from our earlier examples in the lack of any obvious external
stimulus that is closely related to the insight. We shall return to this issue later,
because it bears on our theory.

Hadamard’s theory of scientific insight

Hadamard (1949) gives us a splendid discussion of the phenomenon of insight. In

addition to reviewing numerous instances from the history of science, he identifies
four distinct stages that seem to occur in every documented case of scientific
insight: preparation, incubation, illumination, and verification. These stages and
their characteristics constitute a set of empirical generalizations relating to insight,
and any successful theory must account for their existence.
The preparation phase involves intense effort in attempting to solve a given
problem. In some cases this attempt leads directly to a solution, but for especially
difficult problems one eventually “gives up.” This abandonment constitutes the
entry into the incubation stage, during which one devotes one’s conscious process¬
ing to other issues. Depending on the situation, incubation can last anywhere from
seconds to years, but eventually the solution “proposes itself” during the illumina¬
tion stage, which occurs both unexpectedly and very rapidly. This is the “Aha”
experience that produces exclamations like the “Eureka!” of Archimedes. Howev¬
er, these “leaps of intuition” are not always valid and sometimes lead to “false
insights.” Thus, one still must check the details during the final verification phase.
In addition to describing these four stages and their relation to one another,
Hadamard proposes a theory of insight that gives a major role to unconscious
reasoning. His explanation assumes three levels of the mind that work together
during the process of discovery: the fully conscious, the fringe conscious, and the
unconscious. The first refers to our everyday mode of thought, in which we are
aware of the mental steps we traverse. The unconscious refers to thought processes
that are not available to introspection, of which we are not even aware. The fringe
conscious occupies the gray area between these two extremes, in which we are
aware of ideas but are not focusing on them. One can view this as the “peripheral
vision” area of the mind.
Hadamard’s theory states that the preparation stage involves only conscious
thought. However, the mental activity during preparation serves to “stir up” ideas
relevant to the problem at hand. During the incubation phase, the unconscious mode
takes charge and considers alternative solutions that incorporate the ideas produced
during the earlier preparation. When the unconscious encounters an especially
Computational model of scientific insight 181

promising combination, it deposits the result into the fringe conscious. The mind
seizes on this new idea and experiences the flash of insight as it enters full con¬
sciousness. Finally, one continues in the conscious mode while the result is
checked.
Clearly, most of the action in this theory is occurring at the unconscious level,
and it is natural to ask how this mechanism manages to sift through so many ideas
and distinguish the profitable ones from others. Hadamard argues that the uncon¬
scious is able to generate combinations of ideas that are specific enough to be
fruitful and yet general enough not to miss the solution completely. This process is
likened to firing a shotgun. The pellets spread sufficiently that one does not miss the
target, but not so much that it is useless to aim. Hadamard concludes that great
mathematicians differ from ordinary people in the selective ability of their uncon¬
scious, which lets them generate ideas that are aesthetically pleasing or interesting.

Ohlsson’s restructuring theory

Ohlsson (1984a, 1984b) has proposed a computational model of insight by attempt¬

ing to integrate ideas from gestalt psychology with the problem space framework. In
the gestalt paradigm, every situation was characterized by some structure in the
mind. These structures were influenced by forces that could become unbalanced and
introduce gaps. An unsolved problem was viewed as some situation in which gaps
existed between one’s current state and the goal state. When forces became unbal¬
anced enough, restructuring occurred, and some new configuration was produced.
Gestalt psychologists claimed that these restructuring events were more likely to
occur when the problem solver had carefully analyzed the problem, carefully ana¬
lyzed the goal, and made a series of unsuccessful attempts at solving the problem.
According to the problem space hypothesis, normal problem solving involves a
search througt a problem space. Ohlsson claims that restructuring requires search
through the description space for a problem. That is, restructuring involves finding
a different way to look at the problem, rather than trying to solve the problem in a
straightforward manner. He further assumes that problem solvers are able to “look
ahead” a few steps and that this lets them know when they are near their goal.
When one encounters an impasse, one attempts to view the problem in a different
light. This can lead to a new representation of the problem, which constitutes
restructuring. In some cases, this representational shift leads to a state that is only a
few steps from the goal; the shift combines with the look-ahead ability to produce a
flash of insight.
As an example, Ohlsson presents the problem shown in Figure 7.1, in which one
must compute the sum of the areas of a square and an overlapping parallelogram.
The straightforward solution is to calculate the area of the square and the area of the
parallelogram (which requires calculating the base of the parallelogram) and add
them together. Most people do not know the formula for the area of a parallelogram
182 P. Langley and R. Jones

en AB = a and AG = b.

and thus cannot solve the problem with the information provided. This causes an
impasse, and this in turn leads to restructuring.
One such alteration lets the problem solver view the picture as two overlapping
triangles (DCE and GAB). Given this representation, one can calculate the areas of
the two triangles and add the results. These operations are simple, because the base
and height of the triangles are given in the problem statement. The feeling of insight
might or might not occur in this case, depending on whether or not the problem
solver can look ahead the required three steps.
Another restructuring simplifies the problem even further. If one notices that the
triangles can be “slipped apart” to form a rectangle, then one need only calculate
the area of that rectangle, using the base and height already given. In this case, the
feeling of insight is almost certain to occur, because the goal state is only two steps
away from the initial state in this new space.

Simon’s theory of familiarization and selective forgetting

Simon (1977) has proposed a computational explanation of Hadamard’s four stages

of insight. The theory combines models of human memory with information-pro-
cessing models of problem solving. Research on human short-term memory has
shown that its capacity is severely limited, but it has also shown that this limitation
can be offset by experience. When asked to remember “artificial” input such as
digits or letters, people can hold only about seven symbols in memory at one time.
Computational model of scientific insight 183

However, given sufficient experience in a domain, humans form chunks to describe

regularities in that domain. Simon calls this the process of familiarization. Memory
experiments show that subjects can hold approximately the same number of chunks
in short-term memory, regardless of the complexity of the chunks. Thus, one can
remember around seven letters, seven familiar words, or even seven familiar sen¬
tences. An extreme example is the Gettysburg Address, which is composed of
sentences, which are in turn composed of phrases, which are themselves composed
of certain words, and so forth.
Simon proposes that familiarization occurs during conscious problem solving of
the type that characterizes Hadamard’s preparation stage. As one carries out a
heuristic search through the problem space, one also builds up higher-level struc¬
tures that describe regularities in that space. The goals and states that are generated
during search are held in short-term memory, and these chunks are stored in long¬
term memory. On difficult problems, one can easily get lost and be forced to start
over from the beginning. Meanwhile, one is becoming more familiar with the
structure of the space and its components. Eventually, one may decide that the
problem is too difficult and abandon the efforts. At this point, the structures in
short-term memory will fade rapidly, but the chunks that have been stored in long¬
term memory will remain. Simon terms this process selective forgetting.
Later, one may reexamine the troublesome problem. However, this time one has
a powerful repertoire of chunks available in long-term memory, and these lead the
search down a path quite different from those on earlier occasions. These chunks
may allow one to move directly to the goal, and in some cases this may occur so
rapidly as to produce the experience of illumination. The process of familiarization,
combined with the mechanism of selective forgetting, gives the problem solver a
new approach, thus transforming a difficult problem into a straightforward problem.

Commentary

Let us consider the similarities and differences among Hadamard’s, Ohlsson’s, and
Simon’s theories of the insight process. All assume that Hadamard’s four-stage
model provides a reasonable description of the phenomena, and they concentrate on
explaining the processes that underlie the different stages. Furthermore, all agree
that the preparation and verification stages involve conscious problem solving,
though Ohlsson and Simon give more detail, because they can build on the results of
modem cognitive psychology.
The theories differ in their treatment of the incubation and illumination stages.
Although he does not cast it in quite these terms, Hadamard argues that incubation
involves a search through the space of idea combinations. This search is carried out
by unconscious mechanisms that employ measures of interestingness or elegance,
both to select promising candidates and to decide when a likely solution has been
found. Illumination is secondary in this framework, serving only to notify the
conscious mind of the solution. Most of the interesting action occurs in the uncon-
184 P. Langley and R. Jones

scious during incubation, though the preparation stage also serves to “stir up” the
ideas that are used by the unconscious.
However, developments in cognitive psychology strongly suggest that search of
this kind requires conscious attention. Thus, Simon rejects the notion of an uncon¬
scious that can selectively search large problem spaces of the sort required for many
scientific discoveries. He replaces Hadamard’s unconscious search scheme with
two much simpler unconscious processes: familiarization and selective forgetting.
The first of these occurs during the preparation stage, and the second occurs during
incubation. Together, they clear the way for conscious problem-solving mecha¬
nisms to find a solution during illumination. This stage occurs so quickly because
the chunks acquired in the preparation phase make the search process trivial. This
explanation is much more distributed than Hadamard’s, assigning significant roles
to each stage.
In Ohlsson’s theory, the major action occurs during illumination, when the prob¬
lem solver restructures the problem description so that its solution becomes ob¬
vious. This explanation does not attempt to account for the role of incubation, and
in fact this stage is not even mentioned in the theory. Presumably, Ohlsson would
argue that in some cases restructuring does not occur until some time after an
impasse is reached, but this does not explain what causes restructuring when it does
occur.
Although each of these theories of insight has its attractions, we are not satisfied
that any provides an adequate explanation of the phenomena. Hadamard attributes
powerful search capabilities to the unconscious that contradict the findings of cog¬
nitive psychology. Simon involves the more plausible mechanisms of familiariza¬
tion and selective forgetting, but he does not explain why one returns to the problem
when one does. Finally, Ohlsson posits a restructuring process that generates a new,
simpler problem space. Like Simon’s theory, this framework is consistent with our
knowledge of the human information-processing system, but it does not explain the
incubation stage.
In the following pages we propose an alternative theory of scientific insight that
diverges from the existing theories along a number of dimensions. One difference is
that it does not rely on the problem space hypothesis, as do the approaches of
Simon, Ohlsson, and even Hadamard. Rather, we assume that insight is a memory-
related phenomenon that centers on mechanisms of indexing and retrieval. Another
distinguishing feature of our framework is the central role played by analogy. Given
the importance of this mechanism to our work, we shall diverge slightly to review
some earlier work on the topic.

Research on reasoning by analogy

Looking back on our examples of scientific insight, it becomes apparent that all
involved some form of analogy. Archimedes formed an analogy between his body
Computational model of scientific insight 185

submerged in the bath and the king’s crown submerged in a container of known
volume. Kekule formed a mapping between a snake biting its own tail and the
benzene ring. Finally, Poincare’s insight was based on an analogy between the
Fuchsian transformations and non-Euclidean transformations.
Polya (1945), Sternberg (1977), and others have argued for the importance of
analogy in human cognition. Thus, it would not be surprising to find analogy
occurring in scientific discovery. We shall argue that this mechanism plays an
important role in many (though not necessarily all) cases of insight, and analogy
occupies a central position in our theory of that phenomenon. But before describing
this theory, let us first review some previous work on analogy itself.

Dreistadt’s analogy-based theory of insight

Dreistadt has noted the role of analogy in historical examples of insight. He summa¬
rizes his own theory of insight in the following words (Dreistadt, 1968, p. Ill):
This writer explains insight as occurring when one finds a stimulus pattern (the analogy) in
which parts of the form or structure are like the structure of the problem-situation and the rest
of the structure of this stimulus pattern (the analogy) indicates how to organize the uninte¬
grated materials of the problem or how to recognize the problem by putting the parts that are
out of place into their correct place, or both, thereby completing the whole which is then the
solution of the problem.

In other words, an insight occurs when the problem solver finds some similarity to
the current problem, and this analogy suggests a different view of the problem that
makes its solution clear. In this framework, most of the action occurs during the
illumination stage, which involves the discovery of a suitable analogy.
To test this hypothesis, Dreistadt (1969) performed a number of experiments to
determine the influence of analogies and incubation periods on subjects’ ability to
solve problems. One group of subjects was given 20 min in which to solve a set of
tricky problems. A second group was given the same amount of time to solve the
same problems, but was also presented with pictures that contained analogical hints
to help find the solution. However, this group was not told the purpose of the
pictures. A third group was allowed 5 min to concentrate on the problem, then was
given an 8-min incubation period (involving a distracting activity), and finally was
given 7 min more to solve the problem. A final group was presented with the
pictorial analogies and was given an incubation period.
Dreistadt measured both the number of correct solutions in each group and the
closeness of their incorrect answers. He also interviewed subjects about their im¬
pressions of the problem-solving task. He found that pictorial analogies signifi¬
cantly aided the solution process, even though subjects were not always aware they
had been given a hint. Incubation alone did not seem to help in problem solving, but
there was some evidence that incubation enhanced the effect of the pictorial analo¬
gies. These results lend credibility to the belief that analogies are important in
scientific insight, and we shall return to this view later in the chapter.
186 P. Langley and R. Jones

Hall’s framework for analogy

Although Dreistadt presented evidence that analogy plays a role during insight, he
did not suggest details for this process. However, several researchers in the fields of
artificial intelligence and cognitive science have described computational models of
analogy. Hall (1986) provides an excellent review of these alternative approaches
and suggests an organizing framework for research on analogy. This framework
includes four components: recognition, elaboration, evaluation, and consolidation.
Reasoning by analogy involves mapping from some existing structure, the
source, onto some new structure, the target. One typically begins with an in¬
complete description of the target. The first step involves retrieving a plausible
source from long-term memory; this is the recognition process. Once a likely source
has been identified, one must evaluate the analogy to ensure that it is reasonable.
Assuming the mapping is acceptable, one then carries over relevant aspects of the
source to fill out the target description; this is the elaboration stage. Finally, for
successful analogies, one may want to store an abstract description in memory to
simplify retrieval in future situations; this is the consolidation process.
We shall use this framework in our discussion of the three particular computa¬
tional models of analogy that we consider here. We should note that much of the
work on analogy focuses on learning tasks, and the consolidation stage plays an
important role in this context. However, our focus is on scientific discovery and
insight, and consolidation seems less relevant for this domain. Also, Hall’s frame¬
work downplays the need to store and index experiences in long-term memory
before recognition/retrieval can occur. We shall include this earlier step in our
treatment of analogy.

Gentner’s structure mapping theory

Gentner (1983) has put forth a structure mapping theory that attempts to distinguish
useful analogies from poor ones. This framework assumes that memory contains
representations of objects linked together by predicates. Some predicates accept
only one argument, whereas others relate two or more arguments. The attribute red
is an example of the former; the relation larger is an example of the latter. Gentner
makes a further distinction between first-order predicates, which relate objects, and
second-order predicates, which relate other predicates. The relation larger is an
example of the first, and cause is an example of the second.
The structure mapping theory claims that single-argument attributes are useful
when noting similarities between two situations, but relations are more important
for drawing analogies. For example, the statement “the X12 star system in the
Andromeda galaxy is like our solar system” involves a similarity, implying that the
X12 star is yellow, hot, about the same size as Sol, and so forth. In contrast, the
statement “the hydrogen atom is like our solar system” involves an analogy. In this
case, we certainly do not mean that the hydrogen atom is hot and yellow, but we do
Computational model of scientific insight 187

(a) planet

Figure 7.2. Creating a representation for the atom from the statement “The atom is like
the solar system.” Higher-order relations are carried over, and simple attributes are
ignored.

mean that certain objects (electrons) revolve around its nucleus, more or less as
planets revolve around the sun.
Gentner’s theory does not address the issue of retrieving or recognizing analo¬
gies, but it does provide criteria for evaluating their quality, and it does suggest
principles for carrying out elaboration. The theory can be summarized by three
mapping rules:
1. Disregard attributes of objects, such as size or color.
2. Try to preserve relations between objects.
3. In deciding which relations to preserve, select those that retain consistency among
higher-order relations.

Gentner refers to the third rule as the systematicity principle. The reasoning behind
this principle is that the best analogies retain the highest-order relations.
As an example, consider the partial representation of the solar system shown in
Figure 7.2(a). If we state that “the atom is like the solar system,” the structure
mapping theory predicts that only those relations presented in Figure 7.2(b) will be
carried over. In this case, the sun corresponds to the nucleus of the atom, and the
188 P. Langley and R. Jones

planet maps into the electron. Notice that none of the sun’s attributes are carried
along, nor is the fact that the sun is hotter than its planet, because this relation is not
involved in the higher-order cause relation.

Winston’s theory of analogy

Winston (1980) has proposed an alternative theory that focuses on different aspects
of the analogical reasoning process. As in Gentner’s framework, memory consists
of objects linked together by relations, and together these form schemas describing
some connected set of events. But Winston provides much more than this; his
theory also addresses the issues of indexing and retrieval.
When a new event or description is stored in memory, it is indexed by the type of
object it contains. For example, if one reads a story about a wicked stepmother and
a beautiful girl, the schema summarizing the story will be indexed through those
concepts. Later, when one reads another story involving a wicked stepmother or a
beautiful girl, one will be reminded of the earlier schema. The actual process is both
more complex and more general than this account suggests. Winston organizes
concepts in “is-a” hierarchies, so that “wicked stepmother” will be stored as a
subtype of “stepmother,” “stepmother” will be stored as a subtype of “parent,”
and so forth.
Thus, a story involving any type of parent might remind the reader of the original
schema, though to a lesser extent. But this extension means that, potentially, any
new experience can remind one of any earlier experience. Winston responds to this
issue by preferring connections that are more discriminating during the retrieval
process. For instance, fewer stories will be indexed by “wicked stepmother” than
by the more general “parent” concept. As a result, a new story containing a wicked
stepmother will be more likely to remind one of an earlier story with a wicked
stepmother than will new stories containing other kinds of parents. This approach
has some similarities to “spreading activation” models of retrieval.
Once a plausible source for the analogy has been established in this manner,
Winston’s model compares all possible mappings between the source and the target
and then evaluates them according to their degree of match. This evaluation process
gives preference to higher-order relations, but it also takes objects into account. The
approach also differs from Gentner’s in that the elaboration process carries over
both relations and attributes. Finally, the method consolidates its analogically based
findings by transforming them into production rules, but this process need not
concern us here. Winston has tested his mechanism in a number of domains,
including story understanding and electric circuits.

Carbonell’s theory of derivational analogy

Carbonell (1986) has explored the use of analogy in problem solving. When one
encounters a completely novel problem, the only choice is to employ weak prob¬
lem-solving methods such as heuristic search or means-ends analysis. However,